tags. After the project summary prose, add a single subsection titled "

How can I help on this project?

" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [LLM: Metatag generates and stores tags that were never configured, and keeps only one of them](https://git.drupalcode.org/project/ai/-/work_items/3586723) State: opened | Updated: 2026-09-11 | Author: Ahmad Khader (Ahmad-Khader) | Assigned: unassigned | Labels: category::bug, priority::normal, state::rtbc Description: ## Summary Two defects in the Metatag automator, both independent of the decoding regression fixed in #3586713 and both invisible until that one is fixed, because until then nothing was stored at all. 1. **Every tag is offered to the model, not the configured ones.** `extraAdvancedFormFields()` renders a textarea for every tag Metatag knows about, so the saved configuration carries a key for all of them. `generate()` builds its "available tags" list from those keys rather than from the ones holding a subprompt. The form is explicit that an empty one means *"Keep empty to not generate anythin… ### [Allow Field Widget Actions to target a specific automator 2.x](https://git.drupalcode.org/project/ai/-/work_items/3586657) State: opened | Updated: 2026-09-11 | Author: Artem Dmitriiev (a.dmitriiev) | Assigned: unassigned | Labels: 2026Sprint16, 2026Sprint17, AI Initiative Sprint, ai::inside Description: Port changes from issue https://git.drupalcode.org/project/ai/-/work_items/3571915 to 2.x and 2.0.x branches ### [Review AI issue process and tagging based on recent usage and update docs](https://git.drupalcode.org/project/ai/-/work_items/3550025) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: unassigned | Labels: category::task, documentation, priority::normal, state::needsWork Description: >>> [!note] Migrated issue Reported by: [kristen pol](https://www.drupal.org/user/8389) >>> Problem/Motivation Review the docs and figure out what needs changing based on how we've been using tags/etc: https://project.pages.drupalcode.org/ai/1.1.x/contribute/issue_guidelines Proposed resolution Update docs as necessary Remaining tasks Review, tweak, review, commit > Related issue: [Issue #3540998](https://www.drupal.org/node/3540998) ### [Create AI assistant API session only after the first message](https://git.drupalcode.org/project/ai/-/work_items/3554797) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, AI Innovation, aiAssistantsApi, category::bug Description: >>> [!note] Migrated issue Reported by: [omarlopesino](https://www.drupal.org/user/2853043) Related to !978 >>> Problem/Motivation Currently, when the chatbot is added in the site for anonymous users, a new session is started in order to create the thread. This is necessary, but it can be optimized. Creating a session invalidates reverse proxy caches like Varnish or any kind of CDN, decreasing the site's performance. The session may be only created after the first interaction with the chatbot. Steps to reproduce (required for bugs, but not feature requests) Create an assistant with AI a… ### [Add interactive refinement capability to Field Widget Actions 2.x](https://git.drupalcode.org/project/ai/-/work_items/3586655) State: closed | Updated: 2026-09-11 | Author: Artem Dmitriiev (a.dmitriiev) | Assigned: unassigned | Labels: 2026Sprint16, 2026Sprint18, ai::inside, aiAutomators Description: This issue is for cherry-picking / porting the changes from related issue to 2.x. As in 2.x the way ai_automator works was slightly change, it is not a straightforward cherry-pick. MR to port/cherry-pick https://git.drupalcode.org/project/ai/-/merge_requests/1629 ### [Automator FWA buttons give no user feedback when values don't visibly change](https://git.drupalcode.org/project/ai/-/work_items/3585987) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: unassigned | Labels: 2026Sprint16, 2026Sprint17, AI Initiative Sprint, On hold Description: >>> [!note] Migrated issue Reported by: [petar_basic](https://www.drupal.org/user/3626336) >>> [Tracker] Update Summary: [One-line status update for stakeholders] Short Description: [One-line issue summary for stakeholders] Check-in Date: MM/DD/YYYY Metadata is used by the AI Tracker. Docs and additional fields here. [/Tracker] Problem/Motivation When an editor clicks a Field Widget Action (FWA) button backed by an ai_automator, the AJAX call runs, the throbber appears and disappears, but there is no user feedback confirming that the automator actually ran. This is especially visible on th… ### [Restrict access to the AI API Explorer permission](https://git.drupalcode.org/project/ai/-/work_items/3586733) State: opened | Updated: 2026-09-11 | Author: Marcus Johansson (Marcus_Johansson) | Assigned: Prabhavathi Vanipenta (prabha1997) | Labels: Novice, Quick Win, Release blocker, aiApiExplorer Description: The AI API Explorer is a pure developer tool, but its `access ai prompt` permission is not marked as security-sensitive. Anyone holding it can invoke AI providers directly, and with provider function calling being added it becomes a way to exercise function calls outside the access checks of the features that normally wrap them. That is a meaningful attack surface if the permission is ever granted to a non-developer role. Proposed change in `modules/ai_api_explorer/ai_api_explorer.permissions.yml`: ```yaml access ai prompt: title: 'Permit access to using the AI API Explorer.' description… ### [SetupVdbServer and SetupVdbIndex config actions aren't idempotent, recipe re-apply fails](https://git.drupalcode.org/project/ai/-/work_items/3586729) State: opened | Updated: 2026-09-10 | Author: Abhisek Mazumdar (abhisekmazumdar) | Assigned: Ahmad Khalil (ahmad-khalil-imagex) | Labels: AI Initiative Sprint, ai::inside, category::bug, priority::normal Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !2000 [fix: #3586729 Add existence checks to `SetupVdbServer` and `SetupVdbIndex` config actions](https://git.drupalcode.org/project/ai/-/merge_requests/2000) | State: opened | Author: Abhisek Mazumdar (abhisekmazumdar), 441 diff lines Description: ## Problem/Motivation `SetupVdbServer::apply()` and `SetupVdbIndex::apply()` both call `->create($value)->save()` unconditionally, with no check for whether the target entity already exists: ```php // SetupVdbServer.php, line 111-113 $this->entityTypeManager->getStorage('search_api_server') ->create($value) ->save(); ``` ```php // SetupVdbIndex.php, line 53-55 $this->entityTypeManager->getStorage('search_api_index') ->create($value) ->save(); ``` Any recipe that uses these config actions (e.g. `ai_recipe_content_search_vector`) can't be applied twice on the same site. Re-applying i… ### [Adding Disable Option for AI Automators](https://git.drupalcode.org/project/ai/-/work_items/3508800) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: aiAutomators, category::feature, priority, priority::normal Description: >>> [!note] Migrated issue Reported by: [vakulrai](https://www.drupal.org/user/3524052) Related to !490 >>> Problem/Motivation While working on the AI automator's i felt there is no way we could disable the automator configured on a field , rather than we have to go to edit the field and remove the automator checkbox in order to remove from the automator chain. the drawback i see is admin's have to reconfigure the same set of changes again if they later on enable the same features. Is it a good idea to add this feature and make sure admins can enable and disable the automator's by clickin… ### [Bug — the "Content Entity Seeder" tool generates a JSON Schema with one extra array level and breaks when creating/editing content](https://git.drupalcode.org/project/ai/-/work_items/3586732) State: opened | Updated: 2026-09-10 | Author: Bruno Carvalho (brunocarvalho) | Assigned: unassigned | Labels: category::bug, priority::normal, state::accepted Description: ### **Environment** | **Item** | **Value** | |----------|-----------| | Drupal | 11.4.6 | | PHP | 8.4 | | `drupal/ai` | 1.4.8 | | `drupal/ai_agents` | 1.3.5 (the same `execute()` is present on the `1.4.x` branch / `1.4.0-beta1`) | | Provider / model | OpenAI `gpt-5.4` (via `ai_provider_openai` 1.2.5) | | Tool | `ai_agent:content_entity_seeder` → `Drupal\ai_agents\Plugin\AiFunctionCall\ContentEntitySeeder` | --- ### **Description** `ContextDefinitionNormalizer` wraps the `ContentEntitySeeder`'s `entity_array` parameter in one extra array level (`multiple` + `ComplexToolItems` applied togeth… ### [textToImage() base64-decodes the image payload twice](https://git.drupalcode.org/project/ai/-/work_items/3586730) State: opened | Updated: 2026-09-10 | Author: Sven Decabooter (svendecabooter) | Assigned: unassigned | Labels: aiCoreModule, category::bug, priority::major, state::accepted Description: ## Summary `OpenAiBasedProviderClientBase::textToImage()` calls `base64_decode()` a second time on data it has already decoded, so the `ImageFile` it returns holds corrupt bytes instead of the generated image. The failure is silent: the mime-type guard immediately above inspects the *correct* bytes and passes, and the corrupt bytes are handed out afterwards, so callers get a successful response and write unusable files. The same line also throws away the mime type it just validated in favour of a hardcoded `image/png`. This is a regression. Before commit bc860ac9 ("Let providers select allow… ### [Regression in ai_translate: Layout Builder content is no longer translated](https://git.drupalcode.org/project/ai/-/work_items/3586731) State: opened | Updated: 2026-09-10 | Author: Sven Decabooter (svendecabooter) | Assigned: unassigned | Labels: No Estimation Available, category::bug, priority::critical, translation Description: This issue has been reported in the standalone version of ai_translate module, but the ai_translate module that ships as a submodule within this module is also affected by this: https://git.drupalcode.org/project/ai_translate/-/work_items/3585553 Once a 1.3.x fix for this issue exists in standalone ai_translate, it should probably be backported here as well. Feel free to review / help move the original issue forward, so a fix can be quickly applied here as well. ### [Streaming chat forces plain-text hostname filtering, corrupting URLs inside HTML attributes (e.g., href="")](https://git.drupalcode.org/project/ai/-/work_items/3586624) State: opened | Updated: 2026-09-10 | Author: Ron Raney (rraney) | Assigned: unassigned | Labels: 2026Sprint17, AI Initiative Sprint, On hold, ai::inside Description: **Problem/Motivation** `StreamChatMessageIteratorTrait::getHostnameFilterService()` forces `HostnameFilter::setPlainTextMode(TRUE)` whenever code isn't running inside a PHP Fiber (which is the normal case for typical request-based streaming consumers, e.g. AI CKEditor's Reformat/Modify tools). With plain-text mode forced on, `HostnameFilter::filterText()` skips the HTML-aware filtering branch and instead runs `filterUrlsInPlainText()` — a raw regex substring replace across the entire response text, with no awareness of HTML structure. When AI-processed content contains an `` tag pointing to … --- MERGE REQUESTS (24h) --- ### [Issue #3586657: Allow Field Widget Actions to target a specific automator 2.x](https://git.drupalcode.org/project/ai/-/merge_requests/1999) State: opened | Author: Juan Correa (jucs7) | Branch: 3586657-allow-field-widget-actions-target-specific-automator-2x | 185 diff lines Description: ## Description **Changes:** - `AiAutomatorEntityModifier::saveEntity()` gains an optional `$specificAutomatorId` parameter. When set, processing is filtered to the matching automator config entity ID before any processor runs. - Unlike 1.x, 2.x's `entityHasAutomators()` already returns automators keyed by config entity ID and sorts them with a key-preserving `uasort()`, so the filter here operates directly on that keyed array instead of 1.x's `array_filter()` placed before a key-destroying `usort()`. - Added `testSaveEntityFiltersBySpecificAutomatorId()` and `testSaveEntityReturnsNullForUnma… ### [Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x"](https://git.drupalcode.org/project/ai/-/merge_requests/2002) State: merged 2026-09-11 | Author: Artem Dmitriiev (a.dmitriiev) | Branch: cherry-pick-3d89b17d | 2979 diff lines Description: ## Description ## Testing instructions 1. 2. ## Checklist - [ ] I have linked the related issue in the MR title or description - [ ] I have performed a self-review of my own code - [ ] I have added or updated tests, or explained in the description why this change is not covered by tests - [ ] I have updated documentation for any new or changed functionality - [ ] I have written testing instructions and verified them locally - [ ] I have noted any required post-merge steps (config imports, cache rebuilds, manual changes) - [ ] This MR contains no breaking API or hook changes, or the… ### [feat: #3585987 by petar.basic: Add status message after AI Field Widget Action automator runs.](https://git.drupalcode.org/project/ai/-/merge_requests/1540) State: opened | Author: Petar Bašić (petar_basic) | Branch: 3585987-add-feedback-message | 79 diff lines Description: Closes #3585987 ### [Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x"](https://git.drupalcode.org/project/ai/-/merge_requests/1966) State: merged 2026-09-11 | Author: Juan Correa (jucs7) | Branch: 3586655-add-interactive-refinement-capability-to-field-widget-actions-2x | 2967 diff lines Description: ## Summary Ports the interactive refinement capability for Field Widget Actions Automation from `1.x` (issue #3586468, MR !1629) to `2.x`. As noted in the issue, this was **not** a straight cherry-pick: in `2.x` an `ai_automator` config entity holds an `automator_types` **plugin collection** (each entry its own configured plugin instance) instead of `1.x`'s flat `rule` + `plugin_config`. The port re-implements the same feature against that architecture. ### Changes - **New `Traits/AutomatorFieldWidgetActionTrait.php`** , extracted from `AutomatorBaseAction` so both it and the new `Automato… ### [Draft: fix: Fix AI Settings form to display AI Platforms and Inference Providers](https://git.drupalcode.org/project/ai/-/merge_requests/1901) State: opened | Author: Dezső Biczó (mxr576) | Branch: 3586631-ai-platform-providers-does-not-appear-in-ai-default-settings | 60 diff lines Description: ## Description Closes #3586631 ## TODO - [ ] Test coverage ## Testing instructions 1. 2. ## Checklist - [ ] I have linked the related issue in the MR title or description - [ ] I have performed a self-review of my own code - [ ] I have added or updated tests, or explained in the description why this change is not covered by tests - [ ] I have updated documentation for any new or changed functionality - [ ] I have written testing instructions and verified them locally - [ ] I have noted any required post-merge steps (config imports, cache rebuilds, manual changes) - [ ] This MR co… ### [bug: #3586730 textToImage() base64-decodes the image payload twice](https://git.drupalcode.org/project/ai/-/merge_requests/2001) State: opened | Author: Sven Decabooter (svendecabooter) | Branch: 3586730-texttoimage-base64-decodes-the-image-payload-twice | 189 diff lines Description: ## Description Closes #3586730. `OpenAiBasedProviderClientBase::textToImage()` decoded the `b64_json` payload twice, so the `ImageFile` it returned held bytes that are not an image. The second decode keeps roughly a quarter of the payload, and because the mime-type guard directly above it inspects the *correctly* decoded bytes and passes, the call still reports success — the corruption only surfaces when the file is written and opened. The change is one line in `src/Base/OpenAiBasedProviderClientBase.php`: ```php - $images[] = new ImageFile(base64_decode($image_content), 'image/png', 'gene… --- COMMITS (24h) --- (none) ======================================================================== ## [2] summariseModule:ai_integration_eca:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI Integration - ECA (machine name: ai_integration_eca) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. Issue Title or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no ): "Contributors: Jürgen Haas (jurgenhaas) [1]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with
AI Integration - ECA
then use
,
,
/
, and as needed. Output only the HTML fragment with no surrounding , , or
tags. After the project summary prose, add a single subsection titled "
How can I help on this project?
" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Support media entities for vision models](https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: Marcus Johansson (Marcus_Johansson) | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai::inside Description: >>> [!note] Migrated issue Reported by: [anaconda777](https://www.drupal.org/user/3595801) >>> Problem/Motivation Hi, Currently the ECA Chat task has a text area for prompt. (I am currently using it for categorizing user generated text and it works great) For vision models, there is a need to send media items (like jpg, png etc. files) to the model for categorizing them. So if a content contains a Media field, the file of it should be send to a Model with a prompt. (This by the way works already in Chat Generation Explorer where you can type a prompt and under that is a file upload. ) So a… ### [Stable 1.0.0 release for use in Drupal CMS recipes](https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025) State: closed | Updated: 2026-09-10 | Author: Rajab Natshah (RajabNatshah) | Assigned: unassigned | Labels: category::task, priority::normal, state::fixed Description: Thank you so much Jürgen and Marcus for this module. 934 sites report using it. We added it to our products, in the AI base recipe. We would love to ship it in Drupal CMS recipes. Recipes install with `minimum-stability: stable`, and 1.0.x is rc only, so the install stops: ``` Root composer.json requires drupal/ai_integration_eca ~1, found drupal/ai_integration_eca[dev-1.0.x, 1.0.0-alpha2, ..., 1.0.x-dev (alias of dev-1.0.x)] but it does not match your minimum-stability. ``` You are already at rc5, so a stable 1.0.0 tag would open the door for everyone packaging recipes and products. Happy… ### [Replace token references by roken replacements](https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027) State: closed | Updated: 2026-09-10 | Author: Jürgen Haas (jurgenhaas) | Assigned: unassigned | Labels: category::bug, code, priority::major, state::fixed Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !26 [fix: #3585027 Use token replacement for the token input and mask fields](https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26) | State: merged 2026-09-10 | Author: Jürgen Haas (jurgenhaas), 1355 diff lines Description: In project/ai_integration_eca#3534289 we've changed the file field from not using token references but token replacement. That should also be done everywhere else in the module. --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- - [204a3af8](https://git.drupalcode.org/project/ai_integration_eca/-/commit/204a3af86dbffeb4c36e38d83fd1dc21efab21df) fix: #3585027 Use token replacement for the token input and mask fields — Jürgen Haas (2026-09-10T13:07:04.000+00:00) ======================================================================== ## [3] summariseModule:ai_provider_openai:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: OpenAI Provider (machine name: ai_provider_openai) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. Issue Title or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. Do not name any individual people in this section. No one was active enough this period to list. Format your response as an HTML fragment. Start with
OpenAI Provider
then use
,
,
/
, and as needed. Output only the HTML fragment with no surrounding , , or
tags. After the project summary prose, add a single subsection titled "
How can I help on this project?
" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Add missing OpenAI TTS voices to api_defaults.yml](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3522771) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [dan_metille](https://www.drupal.org/user/574844) Related to !32 >>> Thank you for your great work on this module! I noticed that the current list of OpenAI text-to-speech voices in api_defaults.yml only includes 6 options, while the official OpenAI API now supports 11 voices. I've created a patch (attached) that adds the missing voices: ash, ballad, coral, sage, and verse. This ensures all currently available OpenAI TTS voices can be used with the module. ### [Add test coverage for Whisper (speechToText) and TTS (textToSpeech)](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [alekas](https://www.drupal.org/user/726268) Related to !83 >>> Problem/Motivation The speechToText() (Whisper) and textToSpeech() (TTS) operation overrides in OpenAiProvider currently have zero test coverage. The existing 44 tests only cover Chat, streaming, tool calls, the File API and moderation. Any regression in the audio operations (multipart payload serialization, response parsing, error mapping) would go unnoticed. Proposed resolution Add 43 tests using a real openai-php/client over a Guzzle MockHandler transport — no live API calls, f… ### [Support batched embeddings](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568659) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: 2026Sprint14, 2026Sprint15, AI Initiative Sprint, ai inside Description: >>> [!note] Migrated issue Reported by: [paulsheldrake](https://www.drupal.org/user/1350686) Related to !59 >>> Update the module to support batches for embeddings so they can be inserted into VDBs faster Works with these patches: AI Milvus ### [Moderation endpoint has hardcoded model](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [a.dmitriiev](https://www.drupal.org/user/3235287) >>> Problem/Motivation The method moderationEndpoint is using hardcoded 'omni-moderation-latest' model. This method was there before the moderation operation type was introduced in AI Core module. Now it should of course use the model, that is selected for this operation type. Of course there should be a check that the selected provider is OpenAI. The hardcoded model should be used as a fallback. ### ["Selected API Key not working" after update to AI 1.3.0](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3579428) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, category::bug, code, priority::major Description: >>> [!note] Migrated issue Reported by: [aschiwi](https://www.drupal.org/user/107667) Related to !70 >>> Problem/Motivation This might be a problem caused by the AI module instead of OpenAI Provider but I'm encountering it on an OpenAI Provider config page: I was adding the AI module to an existing site and could not add an OpenAI API key because at admin/config/ai/providers/openai it says this when selecting the valid key "The selected API key is not working. Please double-check the correct API key was entered and that it has credit(s) available.". I tried several things and was confuse… ### [Add reasoning output](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3574226) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !68 >>> Problem/Motivation With the issue https://git.drupalcode.org/project/ai/-/work_items/3546573 we are adding the reasoning summaries as part of the output, so you can see what the model did reason. OpenAI will work as test bed to see that this is working. Steps to reproduce Proposed resolution Introduce reasoning in none-streaming mode. Introduce reasoning in streaming mode if its available while streaming. Remaining tasks User interface changes API changes Data model changes >… ### [Change the way that system prompt is being added in 1.2.x](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3573133) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: Novice, category::task, code, priority::normal Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !65 >>> Problem/Motivation The way historically we set a system prompt was due to the LLM being the main focus at the time, but it was obvious later that the system prompt should be set on the ChatInput instead. This is documented here: https://project.pages.drupalcode.org/ai/1.2.x/developers/develop_third_party_module/#setting-chat-system-messages We should switch over so the OpenAI Provider does this. Proposed resolution Switch from $this->chatSystemRole to fetching it from the i… ### [Add AI usage checkboxes to ai_provider_openai issue template](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568620) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, miscellaneous, priority, priority::major Description: >>> [!note] Migrated issue Reported by: [kristen pol](https://www.drupal.org/user/8389) >>> [Tracker] Update Summary: [One-line status update for stakeholders] Check-in Date: MM/DD/YYYY Additional Collaborators: Metadata is used by the AI Tracker. Docs and additional fields here. [/Tracker] Problem/Motivation Parent: https://git.drupalcode.org/project/ai_initiative/-/work_items/3568605 Issues must declare their AI usage, or they violate DA policies as noted in: Abuse of the Contribution Credit system Proposed resolution Update the issue template with checkboxes from the parent issue. Targ… ### [Add AI 2.x / Search AI compatible OpenAI embedding provider (2.x branch)](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3564086) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [mbatterton](https://www.drupal.org/user/1574576) Related to !57 !58 >>> AI framework 2.x introduces new provider and capability interfaces, which are not compatible with the existing 1.2.x OpenAI provider. This issue introduces a new 2.0.x development branch that: - Implements EmbeddingProviderInterface for AI 2.x - Is compatible with search_ai 2.x - Supports OpenAI embedding models - Leaves 1.2.x fully intact for AI 1.x users This is an embeddings-only provider to support AI Search and vector indexing use cases. AI 2.x separates capabilities expli… ### [Support File Handling](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3548526) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, AI Product Development, category::feature, code Description: >>> [!note] Migrated issue Reported by: [b_sharpe](https://www.drupal.org/user/2512258) Related to !79 !53 >>> Problem/Motivation See https://platform.openai.com/docs/api-reference/files/create Once https://git.drupalcode.org/project/ai/-/work_items/3546954 lands, we can add the ability to upload files and use them for various purposes. Eventually this could be added automatically in some cases (chats with history, fine-tuning). > Related issue: [Issue #3546954](https://www.drupal.org/node/3546954) ### [Remove OpenAI SDK dependency and extend it from Drupal AI core module](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531260) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [gxleano](https://www.drupal.org/user/3591999) Related to !44 >>> Problem/Motivation Currently, OpenAI provider is implemented using the OpenAI SDK directly, with only minor variations. This approach has led to significant code duplication (reused by other providers), especially evident when implementing new features like function calling. In such cases, nearly identical logic had to be added in multiple places, making the codebase harder to maintain, test, and evolve. To address this, we already have an AiProviderClientBase class as a common founda… ### [Handle pdf files in chat](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3529603) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [bircher](https://www.drupal.org/user/1344166) Related to !42 >>> Problem/Motivation Openai can handle pdfs in some models: https://platform.openai.com/docs/guides/pdf-files?api-mode=chat Steps to reproduce Attach a pdf to a chat. Proposed resolution Handle PDFs attached to chat. Remaining tasks adjust the messages when pdfs are attached to chat. User interface changes none API changes none Data model changes none > Related issue: [Issue #3529601](https://www.drupal.org/node/3529601) ### [Version of modules can go out of alignment](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519419) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [mrdalesmith](https://www.drupal.org/user/2612656) Related to !29 >>> Problem/Motivation This is a duplicate of https://www.drupal.org/project/ai_agents/issues/3518506#comment-16067498 to resolve the issue for the main provider in use by people whilst a more generic solution is come up with. Description copied from that issue. In theory you will be able to update the modules in the following order, because of the dependency tree: * AI 1.1.x * AI Agents 1.1.x * Some provider still on 1.0.x This leads to the provider not having any function calling av… ### [Abstract token usage support](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519302) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [aspilicious](https://www.drupal.org/user/172527) Related to !30 >>> https://www.drupal.org/project/ai/issues/3517618 > Related issue: [Issue #3524435](https://www.drupal.org/node/3524435) > Related issue: [Issue #3526936](https://www.drupal.org/node/3526936) > Related issue: [Issue #3526978](https://www.drupal.org/node/3526978) > Related issue: [Issue #3517618](https://www.drupal.org/node/3517618) ### [Usage stats in streaming chat need conversion to an array.](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486224) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [shaundychko](https://www.drupal.org/user/475828) Related to !3 >>> Problem/Motivation \Drupal\ai\OperationType\Chat\StreamedChatMessage::__construct expects the third parameter to be an array. However, when a streaming chat is created with usage included, such as with the following example configuration: /** @var \Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider $provider */ $provider->setConfiguration([ 'stream_options' => [ … ### [Log flagged prompts](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3507407) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::rtbc Description: >>> [!note] Migrated issue Reported by: [mindbet](https://www.drupal.org/user/1703106) Related to !21 >>> Problem/Motivation When loading content in bulk to a vector database, certain paragraphs may be flagged by moderation. The interface does not show exactly what content caused the issue so that it can be addressed. Proposed resolution Add code similar to this to OpenAiProvider.php: (lines 626-628) if (!empty($response['results'][0]['flagged'])) { $moderation_alert = 'This prompt was flagged by the moderation model: ' . $prompt; \Drupal::logger('ai_provider_openai')->… ### [Moderation log overview](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3516044) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [jonas139](https://www.drupal.org/user/2873401) Related to !26 >>> Problem/Motivation When using moderation an error will be thrown when a prompt is flagged. In issue https://www.drupal.org/project/ai_provider_openai/issues/3507407 there's already a logging mechanism but the error is still thrown so it's not very user friendly. It seems better to not throw an error but log the flagged prompt in a separate overview that can be used to track down the related entity (if there is any) so a content manager can review the flagged prompt and rewrite the pi… ### [Streaming chat responses gives Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools()](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3515885) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [cellear](https://www.drupal.org/user/599892) Related to !27 >>> Problem/Motivation From fresh install of Drupal CMS (1.1x-dev) from the Chat Explorer (admin/config/ai/explorers/chat_generator) with Streamed checked, the following error appears: The website encountered an unexpected error. Try again later.Error: Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools() in Drupal\ai_api_explorer\Plugin\AiApiExplorer\ChatGenerator->getResponse() (line 306 of modules/contrib/ai/modules/ai_api_explorer/src/Plugin/AiApiE… ### [Allow skipping of moderations for some embeddings (not all)](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3510601) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, needs forward port, priority::normal Description: >>> [!note] Migrated issue Reported by: [wouters_f](https://www.drupal.org/user/721548) Related to !23 >>> Problem/Motivation Allow the skipping of moderations for some requests, not all. > Related issue: [Issue #3510599](https://www.drupal.org/node/3510599) ### [Bug: Key validation and Rate Limit tester are bypassing host variable; causing Auth errors or Network errors, etc.](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3496412) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [jan kellermann](https://www.drupal.org/user/371731) Related to !43 >>> Problem/Motivation We try to use an openAI compatible API. In OpenAiProvider the config-item `host` is supported - see code. But in OpenAiConfigForm the OpenAi-Client is loaded without custom host - see code so a choise of models is not possible. Proposed resolution Add UI for Host Use \OpenAI::factory() like here. > Related issue: [Issue #3531134](https://www.drupal.org/node/3531134) ### [Selecting "o1" model on automator produces error: "Unsupported parameter: 'max_tokens' is not supported"](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3508042) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, needs backport to 1.1.x, priority::normal Description: >>> [!note] Migrated issue Reported by: [alfthecat](https://www.drupal.org/user/495982) Related to !28 !22 >>> Problem/Motivation Hi, I'm using an automator on a list field and when I switch it to use the "o1" model instead of my previously selected "40o" model under advanced settings, I get the following errors in the log and the field does not get automated/interpolated. "Error invoking model response: Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead." "A general error happened why trying to interpolate, message Error invoking mod… ### [Error with OpenAI o1 models](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489201) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [merilainen](https://www.drupal.org/user/304437) >>> Problem/Motivation I tested the new Open o1 models in AI Chat Explorer, got an error: The following error occurred while processing your request:Error invoking model response: Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.I tied also without max_tokens value but got the same error. ### [Moderation reads confidence scores instead of the actual violation flags, so almost every category shows as flagged](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [rduterte](https://www.drupal.org/user/3809153) >>> Problem/Motivation OpenAI's moderation endpoint returns two different things per category: a confidence score (a number, always present, usually small and non-zero even for harmless text) and a flag (a true or false stating whether that category actually tripped the moderation threshold). This provider's own moderation method reads the confidence scores and stores them as if they were the flags. Since PHP treats any non-zero number as true, and the scores are almost never exactly zero, every categor… ### [Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat()](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [kaszarobert](https://www.drupal.org/user/3578328) >>> Problem/Motivation Using 1.3.0-beta2 throws an error for all calls with the latest stable ai module 1.4.8: Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat() (line 319 of /app/web/modules/contrib/ai_provider_openai/src/Plugin/AiProvider/OpenAiProvider.php)#0 [internal function]: Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat(Object(Drupal\ai\OperationType\Chat\ChatInput), 'gpt-5.6-terra', Array)#… ### [Add support for fast mode](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [scott_euser](https://www.drupal.org/user/3267594) >>> Problem/Motivation A good chunk of the models support fast mode (seems roughly 2x the cost, for 2.5x the response speed) Steps to reproduce See https://developers.openai.com/api/docs/pricing?latest-pricing=fast Proposed resolution Consider how to support; maybe the providers list needs to return them as separate items e.g. gpt 5.4 mini, gpt 5.4 mini (fast mode) as it does not seem desireable to have an all-or-nothing configuration on the OpenAI provider settings made Remaining tasks Decide/agree … ### [Unable to do GPT 5.4 calls](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [jesmaster](https://www.drupal.org/user/1494544) >>> I have my model set to 5.4-nano and it was working fine until this recent update. Now I get "Function tools with reasoning_effort are not supported for gpt-5.4-nano in /v1/chat/completions." If switch to 5.6 it works fine. Most likely related to https://www.drupal.org/project/ai_provider_openai/issues/3613982 ### [How to enable type: web_search in Chat?](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3613029) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::support, documentation, priority::normal, state::postponed Description: >>> [!note] Migrated issue Reported by: [anybody](https://www.drupal.org/user/291091) >>> Postponed on https://git.drupalcode.org/project/ai/-/work_items/3586616 Problem/Motivation In a project we're using Drupal AI for research, including web sources. We're listing different sources to search for in our prompt and we're using OpenAI Provider. I was a bit curious if AI really searches the web for the results (and I'm still) and I found https://developers.openai.com/api/docs/guides/tools-web-search Can someone tell if type: web_search is required to be enabled for AI looking into current r… ### [Automated Drupal 12 compatibility fixes for ai_provider_openai 1.3.x-dev](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3606117) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: Drupal 12 compatibility, ProjectUpdateBotD12, category::task, code Description: >>> [!note] Migrated issue Reported by: [project update bot](https://www.drupal.org/user/3644742) Related to !77 >>> Problem/Motivation Hello project maintainers, This is an automated issue to help make this module compatible with Drupal 12. Changes will periodically be added to this issue that remove deprecated API uses. To stop further changes from being posted, change the status to anything other than Active, Needs review, Needs work or Reviewed and tested by the community. Alternatively, you can remove the "ProjectUpdateBotD12" tag from the issue to stop the bot from posting updates. … ### [Create a skill to setup the provider](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3605812) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, AI Innovation, category::feature, code Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !82 >>> Description We want an agent skill (under ai_provider_openai/.agents/skills/) that sets up the OpenAI provider from an API key that already exists on the system as an environment variable. The skill must work without ever exposing the key's value: it checks only that the variable is set, then wires it in through the Key module's environment key provider so the secret never lands in configuration. This skill does not use the Tool API — it is driven entirely through drush … ### [Automated Drupal 12 compatibility fixes for ai_provider_openai 1.2.x-dev](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3594765) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: Drupal 12 compatibility, ProjectUpdateBotD12, category::task, code Description: >>> [!note] Migrated issue Reported by: [project update bot](https://www.drupal.org/user/3644742) Related to !75 >>> Problem/Motivation Hello project maintainers, This is an automated issue to help make this module compatible with Drupal 12. Changes will periodically be added to this issue that remove deprecated API uses. To stop further changes from being posted, change the status to anything other than Active, Needs review, Needs work or Reviewed and tested by the community. Alternatively, you can remove the "ProjectUpdateBotD12" tag from the issue to stop the bot from posting updates. … ### [Opt-in to cache embedding lookups](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3552523) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::postponed Description: >>> [!note] Migrated issue Reported by: [scott_euser](https://www.drupal.org/user/3267594) Related to !55 >>> Problem/Motivation https://git.drupalcode.org/project/ai/-/work_items/3552522 provides trait method to cache embeddings lookups. Steps to reproduce Embeddings method does not use it Proposed resolution Use the new caching option Remaining tasks Decide 2.0.x branch vs method_exists check User interface changes N/A API changes N/A Data model changes N/A ### [Turn off moderation if omni-moderation-latest is not available](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3534855) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !49 >>> Problem/Motivation When working with a company account in OpenAI, there is the possibility to only allow certain models. This means that you can turn off the moderation models. This is very bad practice since the accounts can get banned, but we have one reported instance of such an account. We need the OpenAI model to still work in these instances and people might not know that we do the opinionated thing of turning on moderation by default. We need to inform when this happens… ### [Add gpt-image-1 image to image capabilitites](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531289) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !45 >>> Problem/Motivation We have added an image to image operation type in https://git.drupalcode.org/project/ai/-/work_items/3531212, meaning that we will be able to take an image and defined parameters and use them and generate an image. OpenAI gpt-image-1 can do this. We should implement it. Steps to reproduce Proposed resolution Implement the ImageToImage operation type with gpt-image-1. Remaining tasks User interface changes API changes Data model changes ### [Warning appears while using automators for image type field "Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() "](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3504849) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::postponed Description: >>> [!note] Migrated issue Reported by: [vakulrai](https://www.drupal.org/user/3524052) >>> Problem/Motivation When i enabled image automator to generate image and selected "default Image to text model" the node save was triggering warnings and not able to generate and save the image in the field. Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() (line 404 of /var/www/html/docroot/modules/contrib/ai_provider_openai/src/Plugin/AiProvider/OpenAiProvider.php) Steps to reproduce Enable and configure the Openai module add a… ### [Add event triggering for streaming](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489846) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !48 >>> Problem/Motivation Add the methods that makes logging possilbe of responses on streaming events from here: https://www.drupal.org/project/ai/issues/3489690 Steps to reproduce Proposed resolution Remaining tasks User interface changes API changes Data model changes ### [Improve documentation on setting key](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486487) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, documentation, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [jvdkolk](https://www.drupal.org/user/562054) >>> The section 'Post-Installation' states: "Set up an API key using the Key module with your OpenAI credentials." Nevertheless, on admin/config/system/keys/add there is a section 'Type settings' - Key type (4 types) - Provider settings - etc. What to set there? I also had to clear the cache before my key showed up on admin/config/ai/providers/openai Proposed text "Set up an API key using the Key module with your OpenAI credentials. Under 'Key type', choose 'Authentication'. Under 'Key provider' choose '… ### [Enable Drupal AI Module Functionality for Environments Behind a Proxy](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3481664) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, proxy Description: >>> [!note] Migrated issue Reported by: [olegiv](https://www.drupal.org/user/3268706) >>> Problem/Motivation The Drupal AI module currently does not support environments where requests must be routed through an HTTP/HTTPS proxy. Many organizations use proxies to secure and manage external communications, especially in development and corporate environments. Without proper proxy support, external API calls made by the module may fail, rendering the module unusable in these environments. Steps to reproduce 1. Set up Drupal in a local or corporate environment where internet access is restrict… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [4] summariseModule:ai_provider_aws_bedrock:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AWS Bedrock Provider (machine name: ai_provider_aws_bedrock) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. Issue Title or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no ): "Contributors: Besart Bytyqi (besart)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with
AWS Bedrock Provider
then use
,
,
/
, and as needed. Output only the HTML fragment with no surrounding , , or
tags. After the project summary prose, add a single subsection titled "
How can I help on this project?
" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Reasoning models return an empty message, and model-specific parameters are rejected by Converse](https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/work_items/3594715) State: opened | Updated: 2026-09-10 | Author: Besart Bytyqi (besart) | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::needsReview Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !19 [Issue #3594715: Read the first content block that carries text, and route model-specific parameters to additionalModelRequestFields](https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19) | State: opened | Author: Besart Bytyqi (besart), 50 diff lines Description: ## Issue summary ### Problem 1: the response is empty for any reasoning model `chat()` reads the answer from content block zero: ```php $message = new ChatMessage($response['output']['message']['role'], $response['output']['message']['content'][0]['text'] ?? ""); ``` Reasoning models put their reasoning in its own content block, and it arrives first. Block zero holds `reasoningContent`, not `text`, so the `?? ""` fires and the provider returns an empty `ChatMessage`. Nothing errors, so the caller sees a successful call that produced no text. That `?? ""` came from #3517795, where block ze… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [5] summariseModule:ai_provider_azure:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Microsoft Azure AI (machine name: ai_provider_azure) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. Issue Title or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. Do not name any individual people in this section. No one was active enough this period to list. Format your response as an HTML fragment. Start with
Microsoft Azure AI
then use
,
,
/
, and as needed. Output only the HTML fragment with no surrounding , , or
tags. After the project summary prose, add a single subsection titled "
How can I help on this project?
" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Structured output: an invalid JSON schema is silently accepted instead of throwing an exception](https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013) State: opened | Updated: 2026-09-10 | Author: Marcus Johansson (Marcus_Johansson) | Assigned: unassigned | Labels: category::bug, code, priority::normal, v1.2.x-dev Description: ## Summary Structured output calls do not fail when the supplied JSON schema is itself invalid. A schema whose `required` array lists a key that is not defined in `properties` is forwarded to the Azure API and the call succeeds, so the caller never learns that the schema is broken. The schema below should not be usable — `description` is listed as required but never defined, while `caption` is defined but never required: ```json { "type": "object", "properties": { "alt": { "type": "string", "description": "Concise alt text for accessibility (max 125 characters)" }, … --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [6] summariseModule:ai_vdb_provider_postgres:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Postgres VDB Provider (machine name: ai_vdb_provider_postgres) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. Issue Title or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no ): "Contributors: Sergiu Nagailic (Nikro), Zeeshaan khann (zeeshan_khan)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with
Postgres VDB Provider
then use
,
,
/
, and as needed. Output only the HTML fragment with no surrounding , , or
tags. After the project summary prose, add a single subsection titled "
How can I help on this project?
" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Add batch embeddings](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3610509) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: 2026Sprint14, 2026Sprint16, AI Initiative Sprint, ai inside Description: >>> [!note] Migrated issue Reported by: [nikro](https://www.drupal.org/user/576416) Related to !30 >>> Problem/Motivation The AI module is introducing batch embeddings as a first-class capability, so indexing can embed and store multiple chunks per request instead of one at a time. This meaningfully reduces indexing time. See the parent work item #3568648: Batch embeddings for improved indexing performance. Vector database providers are being aligned with this capability so they participate in the batched indexing path. The Milvus provider has already been updated in ai_vdb_provider_milvu… ### [indexItems() throws TypeError on array_diff() because $skipDeleteItemIds is never declared](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3614504) State: opened | Updated: 2026-09-10 | Author: Zeeshaan khann (zeeshan_khan) | Assigned: unassigned | Labels: Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !33 [Issue #3614504: Fix TypeError in indexItems() from undeclared $skipDeleteItemIds property](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33) | State: opened | Author: Zeeshaan khann (zeeshan_khan), 18 diff lines Description: ### Problem/Motivation Indexing any content through a Search API index backed by this provider fails with: TypeError: array_diff(): Argument #2 must be of type array, null given in PostgresProvider.php on line 685 This happens on every call to PostgresProvider::indexItems(), because the property $this->skipDeleteItemIds is read but never declared or initialized anywhere in the class (or in its parent, AiVdbProviderClientBase). PHP treats the undefined property access as null, which array_diff() rejects as an invalid second argument. ### Steps to reproduce Step 1: Set up a Search API serve… --- MERGE REQUESTS (24h) --- ### [feat: #3610509 Add batch embeddings](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30) State: opened | Author: Sergiu Nagailic (Nikro) | Branch: 3610509-add-batch-embeddings | 656 diff lines Description: By: nikro Closes #3610509 --- COMMITS (24h) --- (none) ======================================================================== ## [7] summariseModule:canvas:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Drupal Canvas (machine name: canvas) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. Issue Title or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no ): "Contributors: **** (project_19391_bot_cb443e60680723bba08a5351bf595251), Adam G-H (phenaproxima), Ahmad Khalil (ahmad-khalil-imagex) [1], AKHIL BABU (AkhilBabu), Bálint Kléri (balintbrews), Chandan Singh (chandu7929), dpi, Feliksas Mazeikis (f.mazeikis), Fons Vandamme (f0ns), Georgi Ivanov (givanov), John Locke (freelock), Jordan Graham (COBadger), julio retkwa (julioretkwa), Konstantine Kirkitadze (kotekirkitadze), Lauri Timmanee (lauriii), Matt Hand (matthand), Narendra Singh Rathore (NarendraR), Ted Bowman (tedbow), utkarsh_33 (Utkarsh_33), Wim Leers (wimleers), wotnak". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with
Drupal Canvas
then use
,
,
/
, and as needed. Output only the HTML fragment with no surrounding , , or
tags. After the project summary prose, add a single subsection titled "
How can I help on this project?
" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Dependency Dashboard](https://git.drupalcode.org/project/canvas/-/work_items/3591657) State: opened | Updated: 2026-09-11 | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Assigned: unassigned | Labels: projectManagement Description: This issue lists Renovate updates and detected dependencies. Read the [Dependency Dashboard](https://docs.renovatebot.com/key-concepts/dashboard/) docs to learn more. ## Repository Problems Renovate tried to run on this repository, but found these problems. - ⚠️ WARN: Failed to set branch status - ⚠️ WARN: No github.com token has been configured. Skipping release notes retrieval - ⚠️ WARN: Rate limit exceeded for api.github.com, as no hostRules set for this host. Please set a GITHUB_COM_TOKEN - ⚠️ WARN: No tool releases found. ## Deprecations / Replacements > ⚠️ **Warning** > The foll… ### [[>=11.4.7] Broaden our check to determine if a recipe is being applied, use 11.4.7's new `RecipeRunner::isApplying()`](https://git.drupalcode.org/project/canvas/-/work_items/3591985) State: opened | Updated: 2026-09-11 | Author: Adam G-H (phenaproxima) | Assigned: unassigned | Labels: Needs issue summary update, Needs upstream bugfix, Recipes initiative, category::task Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1554 [Check for the recipe runner](https://git.drupalcode.org/project/canvas/-/merge_requests/1554) | State: opened | Author: Adam G-H (phenaproxima), 86 diff lines Description: Right now, we use a pretty narrow hack to determine if a recipe is being applied: https://git.drupalcode.org/project/canvas/-/blob/1.x/src/ComponentSource/ComponentSourceManager.php?ref_type=heads#L100 As the comment notes, the "config is syncing" flag is inconsistent about recipe application. There's no core API to differentiate a normal config sync from a config sync happening during a recipe. Core _will_ add that API in https://www.drupal.org/project/drupal/issues/3613607, but it's not done yet. So until then, we need this hack. But the hack isn't broad enough; we need to know if a recipe… ### [RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it](https://git.drupalcode.org/project/canvas/-/work_items/3591996) State: opened | Updated: 2026-09-11 | Author: Jakub Piasecki (zaporylie) | Assigned: unassigned | Labels: category::bug, priority::critical, state::needsReview Description: ## Overview Applying a site template recipe fails with: In ContainerBuilder.php line 1104: You have requested a synthetic service ("kernel"). The DIC does not know how to construct this service. `RecipeSubscriber::onApply()` calls `PageVariantMigration::migrateDefaultTheme()`, which calls `createFromPageRegions()`, which installs `canvas_page_template_component` when it is missing: ``` // src/PageVariantMigration.php:131 if (!\Drupal::moduleHandler()… ### [[PP-1] Add logic + docs to not ship Component config entities with Drupal extensions (modules, themes, recipes), and how to do it anyway when there's a good enough reason (pre-populated component trees in Canvas Pages or ContentTemplates)](https://git.drupalcode.org/project/canvas/-/work_items/3565712) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: Adam G-H (phenaproxima), Wim Leers (wimleers) | Labels: DX (Developer Experience), Recipes initiative, category::task, configManagement Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1634 [Add docs/recipes.md: how Canvas Component versions interact with recipes...](https://git.drupalcode.org/project/canvas/-/merge_requests/1634) | State: opened | Author: Wim Leers (wimleers), 158 diff lines Description: > [!NOTE] > Issue summary needs updating. Incorporate #3564135+'s different focus. >>> [!note] Migrated issue Reported by: [wim leers](https://www.drupal.org/user/99777) >>> Overview Ever since Component versions were introduced (#3523841: Versioned Component config entities (SDC, JS: prop_field_definitions, block: default_setting, all: slots for fallback) + component instances refer to versions ⇒ less data to store per XB field row), it became implicitly unsupported to ship an exported Component config entity with some Drupal extension in its config/install or config/optional direct… ### [Allow Canvas Patterns+PageRegions+ContentTemplates provided by Recipes to NOT specify component versions](https://git.drupalcode.org/project/canvas/-/work_items/3571366) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: Matt Glaman (mglaman) | Labels: AI-accelerated, Chicago2026, DX (Developer Experience), Drupal CMS dependency critical Description: >>> [!note] Migrated issue Reported by: [penyaskito](https://www.drupal.org/user/959536) Related to !811 >>> Overview In https://git.drupalcode.org/project/canvas/-/work_items/3567413 we added the infra for updating component instances to the latest (active) version if there is no data impact. That is pretty defensive per nature, so we ensure no data loss. But there are some cases where e.g. contrib should be able to impact that. e.g. In \Drupal\canvas\Hook\ShapeMatchingHooks::mediaLibraryStorablePropShapeAlter we implement hook_canvas_storable_prop_shape_alter for media_library, and for… ### [Canvas AI: Add Playwright coverage for the chat Tools dropdown](https://git.drupalcode.org/project/canvas/-/work_items/3591951) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: Konstantine Kirkitadze (kotekirkitadze) | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1607 [test(AI): #3591951 Canvas AI: Add Playwright coverage for the chat Tools dropdown](https://git.drupalcode.org/project/canvas/-/merge_requests/1607) | State: opened | Author: Konstantine Kirkitadze (kotekirkitadze), 202 diff lines Description: ## Summary https://git.drupalcode.org/project/canvas/-/work_items/3591856 added a Tools dropdown to the dev chat UI. Install `canvas_dev_ai`, open the chat, and the dropdown is there. Nothing covers it in a browser. Add Playwright coverage. ## Current behavior No browser coverage exists for the dropdown. The tool list, the selection, the active tool pill, and the `selected_tool` request parameter are all untested. ## Proposed Cover these assertions: 1. Every tool configured in `canvas_dev_ai.settings` renders in the dropdown, with its label and description. 2. With no tools configured, t… ### [Config in language other than site default language causes an exception](https://git.drupalcode.org/project/canvas/-/work_items/3591850) State: opened | Updated: 2026-09-11 | Author: Ted Bowman (tedbow) | Assigned: unassigned | Labels: category::bug, state::needsWork, translation Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1400 [🤖 bug: #3591850 Use entity langcode, not site default, as default translation](https://git.drupalcode.org/project/canvas/-/merge_requests/1400) | State: opened | Author: Ted Bowman (tedbow), 809 diff lines Description: ### Problem When a Canvas config entity's `langcode` differs from the site default language, the Canvas editor throws exceptions. This happens when `PageRegion` (or `ContentTemplate`) entities are created while one language is the site default, and the site default is changed later. The entities keep their original `langcode`. Two exceptions: 1. `InvalidArgumentException: getTranslation() must not be called with the langcode of the default translation ("en")` — thrown when the preview validates a region with an auto-save draft. `CanvasConfigEntityTranslationsAreValidConstraintValidator` and… ### [Canvas AI: Proper chat history management in the backend](https://git.drupalcode.org/project/canvas/-/work_items/3592032) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: AKHIL BABU (AkhilBabu) | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai Description: ## Summary Before starting, read https://git.drupalcode.org/project/canvas/-/work_items/3591923#note_2377604. The dev chat does not keep the agent's history across turns. Each turn runs as a fresh agent: the stored agent state is deleted when the turn finishes, and the next turn rebuilds its history from the client transcript. ## Why Within one turn the backend owns a complete history — every tool call and every tool result — restored from the stored agent state on each hop. At the turn boundary that state is deleted, and the client transcript that replaces it carries message text only, so… ### [Canvas AI: Prevent the selected Tool from changing during a chat turn](https://git.drupalcode.org/project/canvas/-/work_items/3591952) State: closed | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint17, AI Initiative Sprint, ai, ai::inside Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1604 [#3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn](https://git.drupalcode.org/project/canvas/-/merge_requests/1604) | State: merged 2026-09-11 | Author: Ahmad Khalil (ahmad-khalil-imagex), 741 diff lines Description: ## Summary One chat turn in the dev AI chat runs as several requests to `/admin/api/canvas/ai-dev`, all sharing one `request_id`. Each of those requests carries the selected Tool as `selected_tool`. The Tools dropdown stays interactive throughout, so the user can change or clear the Tool while a turn is still running. The selection must be fixed for the whole turn, in the UI and on the server. Part of https://git.drupalcode.org/project/canvas/-/work_items/3591777. ## Current behavior A turn is not one request. Set the component agent as the main agent, open the chat with the browser networ… ### [Site template installation fails with Canvas 1.11](https://git.drupalcode.org/project/canvas/-/work_items/3592052) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: Description: ### Steps to reproduce Setup Drupal CMS with ddev ``` mkdir -p my-drupalcms-site && cd my-drupalcms-site ddev config --project-type=drupal11 --docroot=web ddev start ddev composer create-project drupal/cms ddev composer drupal:recipe-unpack ``` Verify that the canvas version is 1.11.0 ``` ddev composer show drupal/canvas ``` Now run the install command ``` ddev drush si recipes/drupal_cms_starter ``` The install fails with the following error message _`You have requested a synthetic service ("kernel"). The DIC does not know how to construct this service.`_ Now downgrade canvas ve… ### [Canvas AI: Unify error handling in `canvas_component_agent`'s tools](https://git.drupalcode.org/project/canvas/-/work_items/3592021) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1626 [3592021: Add AiResponseValidator::formatErrors() to render accumulated tool...](https://git.drupalcode.org/project/canvas/-/merge_requests/1626) | State: opened | Author: Ahmad Khalil (ahmad-khalil-imagex), 978 diff lines Description: Apply the error-handling pattern from https://git.drupalcode.org/project/canvas/-/work_items/3591958 to the tools `canvas_component_agent` uses: the access check throws first, unchanged; every other check accumulates instead of stopping at the first error; and the accumulated errors are reported through `AiResponseValidator::formatErrors()`. See the "Proposed solution" section of https://git.drupalcode.org/project/canvas/-/work_items/3591958 for the full pattern, and `place_components`/`edit_components` (https://git.drupalcode.org/project/canvas/-/merge_requests/1584) for the reference implem… ### [Remove Canvas-only text format guidelines from non-Canvas forms](https://git.drupalcode.org/project/canvas/-/work_items/3592061) State: opened | Updated: 2026-09-10 | Author: John Locke (freelock) | Assigned: unassigned | Labels: category::bug, priority::minor, state::needsReview Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1637 [bug: #3592061 Remove Canvas-only text format guidelines from non-Canvas forms](https://git.drupalcode.org/project/canvas/-/merge_requests/1637) | State: opened | Author: John Locke (freelock), 133 diff lines Description: Canvas’s block and inline HTML guidelines appear on ordinary text-format fields even though those formats cannot be selected there. Only guidelines for the field’s available formats should appear. ## What is happening? On an anonymous comment form, three sets of instructions appear: - Filtered HTML - Drupal Canvas — Block HTML (locked) - Drupal Canvas — Inline HTML (locked) The form has no text-format selector, so Drupal’s guideline-switching JavaScript cannot hide the extra instructions. Canvas removes its exclusive formats from the selector options on non-Canvas forms, but leaves their … ### [ + `Window.postMessage()` to add defense in depth](https://git.drupalcode.org/project/canvas/-/work_items/3527515) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: JavaScript, Needs tests, Security improvements, category::task Description: >>> [!note] Migrated issue Reported by: [hooroomoo](https://www.drupal.org/user/3688872) >>> Overview https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/iframe#sandbox - Use the sandbox attribute on our current iframe usages - Convert any existing cross-frame communications to use postMessage() When the embedded document has the same origin as the embedding page, it is strongly discouraged to use both allow-scripts and allow-same-origin, as that lets the embedded document remove the sandbox attribute — making it no more secure than not using the sandbox attribute a… ### [Canvas AI: Update orchestrator for content-only / page-variant chrome](https://git.drupalcode.org/project/canvas/-/work_items/3592060) State: opened | Updated: 2026-09-10 | Author: Jordan Graham (COBadger) | Assigned: Jordan Graham (COBadger) | Labels: Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1636 [Draft: task(AI): #3592060 Canvas AI: Update orchestrator for content-only / page-variant chrome](https://git.drupalcode.org/project/canvas/-/merge_requests/1636) | State: opened | Author: Jordan Graham (COBadger), 150 diff lines Description: Problem: After !1624, page/template builders only fill content and refuse site chrome. The orchestrator still describes and routes as if they build headers/footers. **In scope (orchestrator install YAML + post-update to reimport it):** - Tool blurb: drop “including headers, footers…” - Routing: don’t send “add a header/footer” to page builder as if that places chrome - Capability list / examples that promise header/footer pages - Teach the right story: page variant supplies chrome; builders do body/content only (and what to say when the user asks for a header — e.g. explain variant chrome, d… ### [Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form](https://git.drupalcode.org/project/canvas/-/work_items/3592030) State: opened | Updated: 2026-09-10 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1624 [task(AI): #3592030 Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form](https://git.drupalcode.org/project/canvas/-/merge_requests/1624) | State: opened | Author: Ahmad Khalil (ahmad-khalil-imagex), 1352 diff lines Description: Canvas replaced theme-global regions with page variants. Two things follow from that: - The `content` region is unchanged — a page's own content is still always rendered inside it. - Global regions such as header, footer, and sidebar no longer exist. A page variant now lets a site builder decide what the rest of the page looks like — everything besides the main content area — as a single component tree, created and edited entirely from the Canvas editor UI. `canvas_ai` still has two things left over from the old multi-region model: - The `canvas_page_builder_agent` and `canvas_template_buil… ### [Normalize agent-written prop values against their schema and report values the component source cannot store](https://git.drupalcode.org/project/canvas/-/work_items/3592002) State: opened | Updated: 2026-09-10 | Author: Ahmad Khalil (ahmad-khalil-imagex) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai::inside Description: ### Problem/Motivation Split out of #3591921 (MR !1505) at the maintainer's request, see https://git.drupalcode.org/project/canvas/-/work_items/3591921#note_1979702 `AiResponseValidator::validateComponentStructure()` feeds the model's parsed YAML prop values straight into `clientModelToInput()` and the validation constraints. Two consequences show up in live builds with GPT-5.2: 1. **Equivalent textual shapes are rejected.** The model regularly writes `width: '"90%"'`, `level: "2"` or `section_header: "false"`. The YAML parser keeps the inner quotes / string type, every enum, integer and bo… ### [Canvas AI: Unify error handling in the dev page builder agent's tools](https://git.drupalcode.org/project/canvas/-/work_items/3591958) State: opened | Updated: 2026-09-10 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1584 [Unify error handling in all tools used by the dev page builder agent.](https://git.drupalcode.org/project/canvas/-/merge_requests/1584) | State: opened | Author: AKHIL BABU (AkhilBabu), 1460 diff lines Description: **Scope narrowed from the original description.** This issue now covers only the four tools `canvas_dev_page_builder_agent` uses: `place_components`, `edit_components`, `set_page_value`, `get_component_details`. The remaining `canvas_ai` tools get their own follow-up issues, each pointing back to the proposed solution below instead of repeating it: https://git.drupalcode.org/project/canvas/-/work_items/3592021 (component agent), https://git.drupalcode.org/project/canvas/-/work_items/3592022 (live page builder agent's `SetAIGeneratedComponentStructure`), and https://git.drupalcode.org/project/c… ### [Add component label to settings form](https://git.drupalcode.org/project/canvas/-/work_items/3591934) State: opened | Updated: 2026-09-10 | Author: Andrew Morton (mortona2k) | Assigned: Lauri Timmanee (lauriii) | Labels: Needs design, state::needsReview Description: Sometimes the context of what you're editing can get lost. For example, if the selected component is scrolled out of view or the UI has glitched from a recent action, or delayed update. I've seen this happen when doing things quickly. A simple fix that would help is adding the component label to the top of the settings form. ### [The error message about "collapsed input syntax" could be more helpful](https://git.drupalcode.org/project/canvas/-/work_items/3558318) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [phenaproxima](https://www.drupal.org/user/205645) >>> Overview In certain situations, you can receive this validation error: When using the default static prop source for a component input, you must use the collapsed input syntax. I think this message could be better: You can get this message for prop sources that aren't static, like host-entity-url. This is very confusing and can easily point you in the wrong direction -- I'd be inclined to start hunting for problems in my static prop sources, when in fact the problem is not coming from a static… --- MERGE REQUESTS (24h) --- ### [refactor(Conflict resolution): #3591785 Strengthen AutoSaveManager to generate stable hashes](https://git.drupalcode.org/project/canvas/-/merge_requests/1356) State: merged 2026-08-04 | Author: Feliksas Mazeikis (f.mazeikis) | Branch: 3591785-strengthen-autosave-normalization | 977 diff lines Description: ## Summary This branch consolidates the typed-data normalization logic that `AutoSaveManager` and `ComponentSourceBase` previously duplicated into a single method, `TypedDataHelper::castRawPhpTypes()`. ## What changed and why `AutoSaveManager::normalizeEntity()` (hash source) and `ComponentSourceBase::castRawTypedConfigToPhpTypes()` (component-shape normalization) each contained their own primitive-casting and traversal logic. These are merged into `TypedDataHelper::castRawPhpTypes()`, which already existed for the config typed-data path but was not wired into `toStorableArray()`. During c… ### [Draft: Resolve "Non HTML-validating raw template fails to add Canvas slot wrapper comments"](https://git.drupalcode.org/project/canvas/-/merge_requests/1638) State: opened | Author: dpi | Branch: 3569815-non-validating-raw | 320 diff lines Description: [Description of changes] Fully vibe coded, all vibes. Seems to work, pushed so we can share this for testing internally. Feel free to take the work and continue it though! ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3569815 ### [#3592031: Added padding to the container to fix visual gaps.](https://git.drupalcode.org/project/canvas/-/merge_requests/1600) State: opened | Author: utkarsh_33 (Utkarsh_33) | Branch: 3592031-brandkit-colors-ui-fixes | 194 diff lines Description: [Description of changes] ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3592031 ### [Draft: 🤖 feat: #3592009 Add locale-aware date/time formatting filters (Twig) and functions (JS)](https://git.drupalcode.org/project/canvas/-/merge_requests/1632) State: opened | Author: Narendra Singh Rathore (NarendraR) | Branch: 3592009-datetime-formatting-filters-and-functions | 1916 diff lines Description: [Description of changes] AI-Generated: Yes (Used Opencode (AI assistant, anthropic.claude-sonnet-4-6)). ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3592009 ### [Draft: Resolve "Support mapping multi-valued list fields to array-type component props"](https://git.drupalcode.org/project/canvas/-/merge_requests/1630) State: opened | Author: Chandan Singh (chandu7929) | Branch: 3592057-support-mapping-multi-valued-list-fields-to-array-type-component-props | 488 diff lines Description: This change adds support for mapping compatible multi-valued Drupal list fields to array-type Canvas component props. It updates three parts of Canvas: 1. **Validation** `MultiValuedFieldNotSupportedConstraintValidator` now allows supported list field types instead of rejecting all non-single-cardinality fields. This covers: - `list_string` - `list_integer` - `list_float` 2. **Field suggestion / picker flow** `ApiUiContentEntityReferenceControllers` now allows supported multi-valued scalar fields through the UI filtering logic, instead of excluding all multi-valued fields. 3. *… ### [Improve the Brand Kit Fonts UI](https://git.drupalcode.org/project/canvas/-/merge_requests/1520) State: opened | Author: Lauri Timmanee (lauriii) | Branch: 3591957-improve-the-brand-kit-fonts-ui | 3084 diff lines Description: Co-authored by Claude Code ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3591957 ### [chore(deps): update dependency cspell to v10](https://git.drupalcode.org/project/canvas/-/merge_requests/1273) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/major-cspell-monorepo | 1082 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [cspell](https://cspell.org/) ([source](https://github.com/streetsidesoftware/cspell/tree/HEAD/packages/cspell)) | [`^9.2.0` → `^10.0.0`](https://renovatebot.com/diffs/npm/cspell/9.7.0/10.2.2) | ![age](https://developer.mend.io/api/mc/badges/age/npm/cspell/10.2.2?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/cspell/9.7.0/10.2.2?slim=true) | > ❗ **Important** >… ### [chore(deps): update dependency @eslint/compat to v2](https://git.drupalcode.org/project/canvas/-/merge_requests/1306) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/eslint-compat-2.x | 72 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [@eslint/compat](https://github.com/eslint/rewrite/tree/main/packages/compat#readme) ([source](https://github.com/eslint/rewrite/tree/HEAD/packages/compat)) | [`^1.3.1` → `^2.0.0`](https://renovatebot.com/diffs/npm/@eslint%2fcompat/1.4.1/2.1.1) | ![age](https://developer.mend.io/api/mc/badges/age/npm/@eslint%2fcompat/2.1.1?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confide… ### [chore(deps): update dependency @chromatic-com/storybook to v5](https://git.drupalcode.org/project/canvas/-/merge_requests/1305) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/chromatic-com-storybook-5.x | 13 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [@chromatic-com/storybook](https://github.com/chromaui/addon-visual-tests) | [`^3.2.2` → `^5.0.0`](https://renovatebot.com/diffs/npm/@chromatic-com%2fstorybook/3.2.7/5.3.1) | ![age](https://developer.mend.io/api/mc/badges/age/npm/@chromatic-com%2fstorybook/5.3.1?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/@chromatic-com%2fstorybook/3.2.7/5.3.1?slim=true) | >… ### [fix(deps): update npm minor and patch](https://git.drupalcode.org/project/canvas/-/merge_requests/1238) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/npm-minor-and-patch | 103 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [@axe-core/playwright](https://github.com/dequelabs/axe-core-npm) | [`4.11.1` → `4.13.0`](https://renovatebot.com/diffs/npm/@axe-core%2fplaywright/4.11.1/4.13.0) | ![age](https://developer.mend.io/api/mc/badges/age/npm/@axe-core%2fplaywright/4.13.0?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/@axe-core%2fplaywright/4.11.1/4.13.0?slim=true) | | [@babel/parser](… ### [chore(deps): update mysql docker tag to v8.3](https://git.drupalcode.org/project/canvas/-/merge_requests/1272) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/mysql-8.x | 13 diff lines Description: This MR contains the following updates: | Package | Type | Update | Change | Pending | |---|---|---|---|---| | [mysql](https://hub.docker.com/_/mysql) ([source](https://github.com/docker-library/mysql)) | service-image | minor | `8.0` → `8.3` | `8.4` | > ❗ **Important** > > Release Notes retrieval for this MR were skipped because no github.com credentials were available. > If you are self-hosted, please see [this instruction](https://github.com/renovatebot/renovate/blob/master/docs/usage/examples/self-hosting.md#githubcom-token-for-release-notes). --- ### Configuration 📅 **Schedule**: (UT… ### [Iframe sandbox: drop allow-same-origin, add postMessage defense in depth](https://git.drupalcode.org/project/canvas/-/merge_requests/1610) State: opened | Author: Matt Hand (matthand) | Branch: 3527515-iframe-sandbox-windowpostmessage-to-add-defense-in-depth | 2560 diff lines Description: ## Summary This merge request addresses a moderate security vulnerability by removing `allow-same-origin` from Canvas's sandboxed iframes—a combination with `allow-scripts` that represents "a well-known sandbox bypass." The changes affect two components: the code editor's live preview iframe and extension iframes. ## Code Editor Preview Changes The preview iframe has been restructured to eliminate regenerating `srcDoc` on each keystroke. Instead, it loads a fixed shell document once from a dedicated route with an opaque origin (`sandbox="allow-scripts"` only). Live updates now transmit via … ### [issue #3591965: add condition to just load lib on required theme](https://git.drupalcode.org/project/canvas/-/merge_requests/1529) State: opened | Author: julio retkwa (julioretkwa) | Branch: 3591965-canvas-library-is-loaded-on-admin-ui-even-when-not-requested | 14 diff lines Description: adds condition to just load `canvas/canvas.scoped.admin.css` if canvas_stark theme is the default ## Testing instructions * [ ] Install Drupal with the Gin admin theme active. * [ ] Install and enable the Canvas module * [ ] Create or edit a content type that includes a media library field (e.g. an image field using the Media Library widget). * [ ] Navigate to a node edit page for that content type * [ ] Click the "Add media" button to open the media library widget modal. * [ ] Observe that the Gin secondary toolbar (the bar containing "Search", "Shortcuts" ,"Drupal Canvas ") is displaced up… ### [feat(AI): #3592032 Canvas AI: Keep the agent history across chat turns in the backend](https://git.drupalcode.org/project/canvas/-/merge_requests/1635) State: opened | Author: Konstantine Kirkitadze (kotekirkitadze) | Branch: 3592032-issue-branch | 1821 diff lines Description: \## Problem Each dev-chat turn ran a fresh agent. The stored agent state was deleted when a turn finished, and the next turn was seeded from the client transcript, which carries text only. The agent lost every tool call and tool result of the conversation and repeated work, e.g. re-reading component schemas before every follow-up edit. \## Fix \- \`AiWizardDev\` generates one \`conversation_id\` per mount (the point where the chat history is cleared today) and sends it with every request, next to the per-turn \`request_id\`. \- \`CanvasAiTempStore\` gets a second key space, \`conversation_… ### [chore: #3592007 Allow developer and test tooling to locate the Drupal root when the module checkout lives outside the Drupal tree](https://git.drupalcode.org/project/canvas/-/merge_requests/1573) State: opened | Author: Bálint Kléri (balintbrews) | Branch: 3592007-allow-developer-and-test-tooling-to-locate-the-drupal-root-when-the-module-checkout-lives-outside-the-drupal-tree | 109 diff lines Description: ## Problem See #3592007. ## Proposal Honor `DRUPAL_TEST_DRUPAL_ROOT` as the absolute path to the Drupal web root, matching `@drupal/playwright` in all five problematic locations (see #3592007), keeping the existing searches as the fallback so behavior without the variable is unchanged: 1. `composer-root` resolves the project root from it (using `composer.json` beside or above the web root, the same idea `@drupal/playwright` applies). 1. `ui/cypress.config.js` derives the core directory as `$DRUPAL_TEST_DRUPAL_ROOT/core`; the existing `DRUPAL_ROOT_CORE` override keeps working and takes prec… ### [Sync brand kit colors through canvas.brand-kit.json in the Canvas CLI](https://git.drupalcode.org/project/canvas/-/merge_requests/1586) State: opened | Author: Lauri Timmanee (lauriii) | Branch: 3592019-sync-brand-kit-colors-through-canvasbrand-kitjson-in-the-canvas-cli | 5121 diff lines Description: Co-authored with Claude Code. Closes #3592019 ### [fix(Data model): #3558318 The error message about "collapsed input syntax" could be more helpful](https://git.drupalcode.org/project/canvas/-/merge_requests/1633) State: opened | Author: Georgi Ivanov (givanov) | Branch: 3558318-the-error-message-about-collapsed-input-syntax-could-be-more-helpful | 203 diff lines Description: ## Problem `ValidComponentTreeItemConstraintValidator` reported: > When using the default static prop source for a component input, you must use the collapsed input syntax. Two problems: 1. The violation fires for any prop source, not only static ones. A `host-entity-url` prop source stored as `{"sourceType": "host-entity-url"}` triggers it, because `absolute` defaults to `TRUE` and becomes explicit once collapsed. The message named the wrong cause. 2. "Collapsed input syntax" was defined in no docblock, comment or documentation. ## Solution The message now names the prop at fau… ### [Issue #3592055 by f0ns: Hide auto-created folders whose Components are all disabled](https://git.drupalcode.org/project/canvas/-/merge_requests/1628) State: opened | Author: Fons Vandamme (f0ns) | Branch: 3592055-library-shows-empty-auto-created-folders-for-categories-whose-components-are-all-disabled | 378 diff lines Description: ## What this does `Component::postSave()` created a Folder for every new Component regardless of status. Core-provided Block Components are generated disabled, so each of their categories got a Folder that shows up empty in the Library. Folder config entities are keyed by UUID, so a site template can't target them with config actions. A Folder is now only created for a Component that is **enabled**. That covers both a Component created enabled and one switched on later. It is deliberately *not* the `FALSE → TRUE` transition alone, since SDCs are created enabled and never undergo that transit… ### [fix(CLI Tool): #3592059 Improve Canvas CLI compatibility with Canvas <1.11](https://git.drupalcode.org/project/canvas/-/merge_requests/1631) State: opened | Author: wotnak | Branch: 3592059-improve-canvas-cli-compatibility-with-canvas-111 | 654 diff lines Description: - move the page template support check before authentication - continue syncing pages and content templates when page templates are not support just with page template fields skipped - improve page template unsupported message to suggest using CLI 0.23 on older sites ## Testing instructions - try to push/pull with `@drupal-canvas/cli@0.24.0` on a Canvas 1.10 site it will fail with missing oauth scope errors - rebuild the CLI with the MR changes - try to push/pull on the same site but with the rebuild CLI, it should now display helpful warning about page templates not being supported and pull/… --- COMMITS (24h) --- - [4c5b55b9](https://git.drupalcode.org/project/canvas/-/commit/4c5b55b97b062174f729f7bb3fd12ad11eebddc1) task(AI): #3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn — Ahmad Khalil (2026-09-11T06:23:48.000+00:00) ======================================================================== ## [8] summariseModule:ai_context:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Context Control Center (CCC) (machine name: ai_context) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. Do not name any individual people in this section. No one was active enough this period to list. Format your response as an HTML fragment. Start with <h3>Context Control Center (CCC)</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [CCC Pre-rc1 QA: Context items settings testing](https://git.drupalcode.org/project/ai_context/-/work_items/3586496) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::task Description: ## Problem/motivation Child of #3586460+s. Context item settings control publish-by-default, revisions, subcontext hierarchy, conditional AI decisions, and the Scheduler link when that module is installed. ## Solution Run the attached script `3586496-context-items-settings.md`. Cover publishing defaults, hierarchy off, conditional AI on/off and the parent-decision cap, deep links, and Scheduler. Comment Pass / Fail / Blocked / N/A for each case. File one issue per bug and relate it here and to #3586460+s. [00-shared-setup.md](https://git.drupalcode.org/-/project/191993/uploads/80e34ffdf… ### [Disabling "Create new revisions by default" still creates a new revision](https://git.drupalcode.org/project/ai_context/-/work_items/3586515) State: opened | Updated: 2026-09-11 | Author: tushar1 (Tushar1) | Assigned: unassigned | Labels: Description: ## Description Disabling **Create new revisions by default** does not prevent a new revision from being generated when an existing AI Context Item is edited and saved. ## Steps to reproduce 1. Go to `/admin/config/ai/context/settings/items`. 2. Enable **Create new revisions by default**. 3. Save the configuration. 4. Edit an existing AI Context Item and make a content change. 5. Save the item and verify that a new revision is created. 6. Return to Context Items Settings. 7. Disable **Create new revisions by default**. 8. Save the configuration. 9. Edit the same AI Context Item and … ### [Update context consumer listing configuration summary column](https://git.drupalcode.org/project/ai_context/-/work_items/3586514) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: Kristen Pol (kepol) | Labels: 2026Sprint18, consider for 1.0, state::toBeTriaged Description: to include what subscriptions are set ### [Change Add AI Context Item button to Add Context Item](https://git.drupalcode.org/project/ai_context/-/work_items/3586513) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: Kristen Pol (kepol) | Labels: 2026Sprint18, AI Initiative Sprint, ai::internal, category::task Description: I keep forgetting to do this and it's been on my mental list for many months ### [Accessibility Testing: CCC overview dashboard](https://git.drupalcode.org/project/ai_context/-/work_items/3586434) State: opened | Updated: 2026-09-10 | Author: Kristen Pol (kepol) | Assigned: Kati Vera (kati.vera) | Labels: 2026Sprint18, AI Initiative Sprint, accessibility, ai::inside Description: ## Problem/motivation Parent: #3586433+s _**THIS ISSUE IS RESERVED FOR MEARRA (KATI) - PING KRISTEN IN DRUPAL SLACK YOU WANT TO WORK ON THIS AND AREN'T KATI :smile:**_ _**THIS ISSUE WAS GENERATED WITH GROK AND NEEDS HUMAN REVIEW BEFORE PROCEEDING.**_ This is the first-batch area **#2**. This is an **area audit**, not a confirmed bug. Run the checks below on the Overview dashboard and file each confirmed finding as its own follow-up (one WCAG success criterion per issue). Do not change code from reading the template or from AI review alone. CCC has had keyboard/DOM work and some automat… ### [CCC Pre-rc1 QA: Test context item creation and workflow](https://git.drupalcode.org/project/ai_context/-/work_items/3586491) State: closed | Updated: 2026-09-10 | Author: Kristen Pol (kepol) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::task Description: ## Problem/motivation Child of #3586460+s. Context items are the core CCC content. Editors must be able to create, edit, view, duplicate, revise, moderate, and delete items without data loss or broken listing/view pages. ## Solution Run the attached script `3586491-context-item-creation.md`. Cover add/edit/view/delete/duplicate, Purpose validation, listing, revisions, Global vs scopes, subcontext rules, moderation, scheduling, import, and access. Comment Pass / Fail / Blocked / N/A for each case. File one issue per bug and relate it here and to #3586460+s. [00-shared-setup.md](https://g… ### [CCC Pre-rc1 QA: Context scope testing](https://git.drupalcode.org/project/ai_context/-/work_items/3586492) State: opened | Updated: 2026-09-10 | Author: Kristen Pol (kepol) | Assigned: Tazo Vekua (tazo_vekua) | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::task Description: ## Problem/motivation Child of #3586460+s. Scope settings and per-item scope widgets decide what is eligible for selection. Bugs here silently drop or leak context. ## Solution Run the attached script `3586492-context-scope.md`. Cover the scope overview, enable/disable without wiping stored values, Language unavailable on a single-language site, Site Sections, Entity Types, Roles, Specific Entities, Taxonomy, and consumer subscriptions. Comment Pass / Fail / Blocked / N/A for each case. File one issue per bug and relate it here and to #3586460+s. [00-shared-setup.md](https://git.drupalc… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [9] summariseModule:ai_ckeditor:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI CKEditor (machine name: ai_ckeditor) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Kristen Pol (kepol)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI CKEditor</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections](https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, category::bug, code Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !16 [Issue #3615848: Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections.](https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16) | State: opened | Author: Kristen Pol (kepol), 916 diff lines Description: ## Problem Follow-up to: ai_context#3586510+s `AiRequest::doRequest()` attaches `entity_context` request metadata so subscribers such as AI Context can scope pushed context to the entity being edited. It only works for saved entities. 1. **Unsaved entities have no context at all.** `ai_ckeditor_form_alter()` puts `entity_type` and `id` into drupalSettings. On an add form the id is an empty string, and `validateEntityContext()` returns NULL when the id is empty, so nothing is sent. AI Context's Entity Types scope then cannot match, so a context item scoped to "Content: Pirate post" is never… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [10] summariseModule:ai_dashboard:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI Dashboard (machine name: ai_dashboard) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Rob Loach (robloach) [1]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI Dashboard</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [How often is the ecosystem data updated?](https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221) State: opened | Updated: 2026-09-10 | Author: Michael Anello (ultimike) | Assigned: unassigned | Labels: Description: I'm curious how often the data on https://ai-dashboard-41c651.pages.drupalcode.org/ecosystem/ is updated. thanks, -mike --- MERGE REQUESTS (24h) --- ### [Issue #3585221: Rebuild the ecosystem](https://git.drupalcode.org/project/ai_dashboard/-/merge_requests/34) State: merged 2026-09-10 | Author: Rob Loach (robloach) | Branch: 3585221-rebuild_ecosystem | 4616 diff lines Description: Refreshes ecosystem.json and ecosystem.md (234 modules, 38 recipes as of 2026-09-10), and fixes the crawler to retry raw socket timeouts. --- COMMITS (24h) --- - [dfddf5e5](https://git.drupalcode.org/project/ai_dashboard/-/commit/dfddf5e5b35971cb0c30a08922b2971359c113b4) Issue #3585221: Rebuild the ecosystem — Rob Loach (2026-09-10T16:16:17.000+00:00) ======================================================================== ## [11] summariseModule:field_widget_actions:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Field Widget Actions (machine name: field_widget_actions) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Artem Dmitriiev (a.dmitriiev), Joonas Meriläinen (mErilainen), Nik LePage (niklp) [1]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>Field Widget Actions</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Target element lookup fails for select widgets](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828) State: opened | Updated: 2026-09-11 | Author: Jürgen Haas (jurgenhaas) | Assigned: unassigned | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai::inside Description: FieldWidgetActionBase::getTargetElement() appends the constant FORM_ELEMENT_PROPERTY ('value') to the triggering button's parents when a delta exists, and falls back to the hardcoded path field/widget/0/value otherwise. That only matches widgets whose input is a value child, like plain text fields. Select, radios, checkboxes, tagify and select2 widgets fail: core marks them multiple_values: TRUE, so WidgetBase::form() never builds delta keys, and this module's fieldWidgetSingleElementFormAlter() wraps them in a container with the input under the widget key. The lookup then returns an empty ar… ### [Handle empty "Set widget field value" gracefully (with helpful message)](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578827) State: opened | Updated: 2026-09-11 | Author: Jürgen Haas (jurgenhaas) | Assigned: Shashank Rai (shqshqnk) | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai::inside Description: This got originally reported in project/eca_field_widget_actions#3588905+ by @ultimike We came to the conclusion that this should be handled by the FWA framework, but we can't move the other issue over here. Therefore, opening a new one. ### [Bug when trying to "Remove action"?](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578826) State: opened | Updated: 2026-09-11 | Author: Michael Anello (ultimike) | Assigned: unassigned | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai::inside Description: I have a fresh Drupal 11.4.5 site with AI 1.4.7 and Field Widget Actions 1.4.1 After creating a simple "LLM: text" automator and adding it to a Text (plain, long) field via FWA, it works as expected - no problems there. But, if I return to the "Manage form display" page for the content type and try to remove the FWA from the field's configuration, it doesn't let me. Here's exactly what what I did and what happens: 1. `drush cr` (because, of course I did) 2. Navigate to the "Manage form display" page for the content type. 3. Click to configure the Text field (via its gear icon.) 4. Opened th… ### [Configured actions are invisible in the "Manage form display" summary row](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578830) State: closed | Updated: 2026-09-10 | Author: Nik LePage (niklp) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::feature Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !32 [Add action summary to manage fields ui](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/32) | State: merged 2026-09-10 | Author: Nik LePage (niklp), 45 diff lines Description: Created with the help of AI (CC) When a field widget has one or more actions configured (e.g. "Draft with AI"), that's only visible by opening the widget's settings (the gear icon) and expanding the "Field Widget Actions" details. The collapsed summary row on Manage form display — the same line that shows things like "Number of rows: 8" for a textarea — gives no indication that any actions are attached at all. This makes it easy to lose track of which fields have actions configured, especially on forms with many fields, since you have to open every field's settings individually to check. Dr… ### [TranslatableMarkup returned as string causes fatal error](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578829) State: closed | Updated: 2026-09-10 | Author: Nik LePage (niklp) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::bug Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !33 [#3578829: Cast return value to string in getDescription()](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/33) | State: merged 2026-09-10 | Author: Joonas Meriläinen (mErilainen), 13 diff lines Description: Created with the help of AI Current dev branch (src/FieldWidgetActionBase.php:106-108): ``` public function getDescription(): string { return $this->getPluginDefinition()['description'] ?? ''; } ``` The attribute declares $description as ?TranslatableMarkup (Attribute/FieldWidgetAction.php), so any non-null description returns a TranslatableMarkup object under the string return type — which causes a fatal. The `?? ''` only guards the null/missing case, not the type mismatch. Fix: one-line cast — `return (string) ($this->getPluginDefinition()['description'] ?? '');` --- MERGE REQUESTS (24h) --- ### [Add action summary to manage fields ui](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/35) State: merged 2026-09-10 | Author: Artem Dmitriiev (a.dmitriiev) | Branch: cherry-pick-ea54b3d2 | 45 diff lines Description: (cherry picked from commit ea54b3d201221ef49ee4ce0d9bc245c052c65320) 97097c30 Add action summary to manage fields ui 24ae1493 Fixed coding standards error f1ecf37d Fix coding standards Co-authored-by: Nik LePage ### [#3578829: Cast return value to string in getDescription()](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/34) State: merged 2026-09-10 | Author: Artem Dmitriiev (a.dmitriiev) | Branch: cherry-pick-76fdceae | 13 diff lines Description: (cherry picked from commit 76fdceaed36a5c606bd64ff5185a7c5942945554) 76fdceae #3578829: Cast return value to string in getDescription() Co-authored-by: Joonas Meriläinen --- COMMITS (24h) --- - [ea54b3d2](https://git.drupalcode.org/project/field_widget_actions/-/commit/ea54b3d201221ef49ee4ce0d9bc245c052c65320) Add action summary to manage fields ui — Nik LePage (2026-09-10T13:57:15.000+00:00) ======================================================================== ## [12] summariseModule:ai_translate:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI translate (machine name: ai_translate) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Sven Decabooter (svendecabooter)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI translate</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [HTML markup is corrupted/not preserved when translating formatted text fields (text_with_summary, text_long)](https://git.drupalcode.org/project/ai_translate/-/work_items/3585525) State: opened | Updated: 2026-09-10 | Author: Carlos Busto Capeáns (cbccharlie) | Assigned: unassigned | Labels: automation::suggestionExists, category::bug, needs port to 2.x, priority::normal Description: ## Problem/Motivation When using AI Translate to translate content that contains formatted text fields (`text_with_summary`, `text_long`) with HTML markup (e.g. ``, ``, ``, ``, etc.), the translated output does not preserve the original HTML structure. The AI model either strips, alters, or partially reconstructs the HTML tags, resulting in broken or invalid markup stored in the translated field value. This affects any content type that uses fields with a text format such as "Full HTML" or "Basic HTML" via CKEditor. ## Steps to reproduce 1. Install and configure the `ai_translate` submodul… ### [Regression: Layout Builder content is no longer translated](https://git.drupalcode.org/project/ai_translate/-/work_items/3585553) State: opened | Updated: 2026-09-10 | Author: Sven Decabooter (svendecabooter) | Assigned: unassigned | Labels: category::bug, code, needs backport to 1.3.x, needs port to 2.x Description: ## Summary This occurs since the changes introduced by https://www.drupal.org/sa-contrib-2026-121 `TextExtractor::shouldExtract()` calls `access('view')` on every field, and core's `LayoutSectionItemList::defaultAccess()` returns `AccessResult::forbidden()` unconditionally, so `layout_builder__layout` fails that check for every account, including user 1 with `configure any layout`. The layout field is therefore dropped before `LbFieldExtractor` ever runs, and Layout Builder inline block content is never translated. The check was added in 1.3.2 and 1.4.1; the failure is silent, so translation… --- MERGE REQUESTS (24h) --- ### [Resolve "Regression: Layout Builder content is no longer translated"](https://git.drupalcode.org/project/ai_translate/-/merge_requests/49) State: opened | Author: Sven Decabooter (svendecabooter) | Branch: 3585553-issue-branch-1.4 | 708 diff lines Description: ## Description Closes #3585553. `TextExtractor::shouldExtract()` checks `access('view')` on every field, and core's `LayoutSectionItemList::defaultAccess()` returns `AccessResult::forbidden()` for every account, including user 1. So `layout_builder__layout` never passes that check, the layout field is dropped before `LbFieldExtractor` is reached, and Layout Builder inline blocks are never translated. The translation still saves and reports success, so the failure is silent. Affects 1.3.2 and 1.4.1, with both `layout_builder_at` and `layout_builder_st`. The fix adds `ContainerFieldTextExtrac… --- COMMITS (24h) --- (none) ======================================================================== ## [13] summariseModule:elevenlabs:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: ElevenLabs (machine name: elevenlabs) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Anne Bonham (annebonham)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>ElevenLabs</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Text-to-speech audio is truncated to the first ~5,000 characters because MP3 chunks are joined by raw byte concatenation](https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581) State: opened | Updated: 2026-09-11 | Author: Anne Bonham (annebonham) | Assigned: unassigned | Labels: Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !5 [Issue #3597581 by banoodle: Fix audio truncated to the first chunk when text...](https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5) | State: opened | Author: Anne Bonham (annebonham), 282 diff lines Description: **Problem/Motivation** On a site where I have this enabled, the client noticed that the ElevenLabs-generated audio files were playing only the first 5,000 characters of text, even though the Elevenlabs limit is 10,000 characters.\ \ `ElevenlabsProvider::textToSpeech()` splits text longer than 5,000 characters into\ chunks, requests one MP3 per chunk, and then joins the results. The join is a raw\ byte append:\ \ `audio_binary_data = _NULL_;\ foreach (``audio_files as $audio_file) {$audio_binary_data .= $audio_file->getBinary();}`\ \ Every response from the ElevenLabs API is a complete MP3 fil… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [14] summariseModule:flowdrop:developer ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: FlowDrop (machine name: flowdrop) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a technical developer audience. Focus on: what was merged or shipped, specific bugs fixed, APIs changed, and what is blocking progress. Be specific — mention function names, module names, and MR references where relevant. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Shibin Das [8]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>FlowDrop</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a developer. Suggest 2-3 concrete technical actions a contributor could take right now. Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, check the data for each issue: - If an issue shows a Related MR with state 'opened', do NOT suggest creating a patch or MR — one already exists. Suggest reviewing it instead. - If all Related MRs for an issue are merged, do NOT suggest reviewing them — they are already done. - If an issue is unassigned with no Related MRs, it is a good candidate to pick up. - Only suggest actions that are genuinely still needed given the current state in the data above. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Bare editor/playground page leaks <head-placeholder>, pushing every stylesheet into <body> so UI themes (Drafter, Minimal) and skins never apply](https://git.drupalcode.org/project/flowdrop/-/work_items/3592452) State: opened | Updated: 2026-09-10 | Author: Shibin Das (D34dMan) | Assigned: unassigned | Labels: bug Description: ## Problem `BarePageRenderer` (flowdrop_ui_components) emits the literal tag `` in its HTML template. Drupal core's `HtmlResponseAttachmentsProcessor` only replaces the head placeholder when the response attaches a non-empty `html_head`. The bare page attaches none, so the raw tag is shipped to the browser. An unknown element is not allowed inside ``, so the HTML parser closes `` at that point and everything after it — the ``, the inline reset `` and **every ``** — is parsed into ``, wrapped in a stray `` element. ## Effect The `@flowdrop/flowdrop` library applies a UI theme (Drafter, Mini… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- - [2c459ecc](https://git.drupalcode.org/project/flowdrop/-/commit/2c459eccb3ece34235e0d5e3a4e9bc71556d9282) ci: exclude the vendored WebMCP bridge from eslint, prettier and cspell; add four dictionary words — Shibin Das (2026-09-10T23:58:54.000+02:00) - [49666c5f](https://git.drupalcode.org/project/flowdrop/-/commit/49666c5f14d7138d6e70d880672ab29cc716792f) chore: re-stamp 2.6.0-alpha1 with the chat tool loop — Shibin Das (2026-09-10T23:23:44.000+02:00) - [fca1f985](https://git.drupalcode.org/project/flowdrop/-/commit/fca1f98516e5e8898d12d2581bbe7c16b9aed7a6) docs: CHANGELOG — Unreleased entries for the chat tool loop, the editor surface switches, the bare page head fix and the 2.8.0 bundle — Shibin Das (2026-09-10T23:13:19.000+02:00) - [98717d50](https://git.drupalcode.org/project/flowdrop/-/commit/98717d5026854f8051bf5790f88d3d7ef66c0e2c) fix(flowdrop_chat): #3592451 a tool-heavy turn no longer wedges on the memory cap — Shibin Das (2026-09-10T23:13:19.000+02:00) - [0f351c9c](https://git.drupalcode.org/project/flowdrop/-/commit/0f351c9cedd4a94691ddf6eae12cead1b03d6a0c) test(flowdrop_ui_components): SurfaceSettingsTest no longer installs flowdrop_chat — Shibin Das (2026-09-10T23:13:19.000+02:00) - [e1eb94c2](https://git.drupalcode.org/project/flowdrop/-/commit/e1eb94c2800ed38ed9f2a5351254ed45356ab1ec) chore: bump flowdrop js lib to 2.8.0 — Shibin Das (2026-09-10T23:05:11.000+02:00) - [5a4a2b69](https://git.drupalcode.org/project/flowdrop/-/commit/5a4a2b69b714c56ff31b7b0df399eacbfc311d6d) feat(flowdrop_ui_components): site switches for the Command Console and the AI Assistant; no dead Sync to Cloud — Shibin Das (2026-09-10T16:36:01.000+02:00) - [e8e5dc4c](https://git.drupalcode.org/project/flowdrop/-/commit/e8e5dc4c0265a82f2648176959196414e05672ca) fix(flowdrop_ui_components): #3592452 bare page leaks head placeholder, stylesheets land in body — Shibin Das (2026-09-10T15:45:56.000+02:00) ======================================================================== ## [15] generateTldr:developer ======================================================================== You are an editor distilling a Drupal AI project newsletter into its most important highlights. You are writing for a technical developer audience. Be specific — name modules, merged features, and critical bugs. Read all the module summaries below. Separate the highlights into two categories: - SHIPPED: things that were merged, fixed, released, or completed during this period. - ONGOING: things that are actively in progress, under review, or blocked. Be specific — name the module, what happened, and why it matters. Do not use emoticons or mdashes. Do not include any text outside the two sections. When you mention a specific issue or merge request, add an inline citation using [N] where N is the reference number from the list below. You may cite multiple references per item, e.g. [1][3]. Place citations immediately after the relevant phrase, before any punctuation. Format as two HTML sections. Use exactly this structure (all <li> elements must be inside the <ol>, never outside it): <h4>Shipped</h4> <ol> <li><strong>Title here</strong> — One sentence explanation [1].</li> <li><strong>Another title</strong> — One sentence explanation.</li> </ol> <h4>Ongoing</h4> <ol> <li><strong>Title here</strong> — One sentence explanation [2].</li> </ol> Up to 5 items per section. Do not output any text, tags, or characters outside these two sections. Output only the HTML fragment, no surrounding tags. --- REFERENCE LIST --- [1] LLM: Metatag generates and stores tags that were never configured, and keeps only one of them (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586723 [2] Allow Field Widget Actions to target a specific automator 2.x (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586657 [3] Review AI issue process and tagging based on recent usage and update docs (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3550025 [4] Create AI assistant API session only after the first message (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3554797 [5] Add interactive refinement capability to Field Widget Actions 2.x (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586655 [6] Automator FWA buttons give no user feedback when values don't visibly change (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3585987 [7] Restrict access to the AI API Explorer permission (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586733 [8] SetupVdbServer and SetupVdbIndex config actions aren't idempotent, recipe re-apply fails (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586729 [9] Adding Disable Option for AI Automators (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3508800 [10] Bug — the "Content Entity Seeder" tool generates a JSON Schema with one extra array level and breaks when creating/editing content (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586732 [11] textToImage() base64-decodes the image payload twice (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586730 [12] Regression in ai_translate: Layout Builder content is no longer translated (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586731 [13] Streaming chat forces plain-text hostname filtering, corrupting URLs inside HTML attributes (e.g., href="") (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586624 [14] Issue #3586657: Allow Field Widget Actions to target a specific automator 2.x (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1999 [15] Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x" (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2002 [16] feat: #3585987 by petar.basic: Add status message after AI Field Widget Action automator runs. (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1540 [17] Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x" (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1966 [18] Draft: fix: Fix AI Settings form to display AI Platforms and Inference Providers (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1901 [19] fix: #3586729 Add existence checks to `SetupVdbServer` and `SetupVdbIndex` config actions (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2000 [20] bug: #3586730 textToImage() base64-decodes the image payload twice (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2001 [21] Support media entities for vision models (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289 [22] Stable 1.0.0 release for use in Drupal CMS recipes (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025 [23] Replace token references by roken replacements (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027 [24] fix: #3585027 Use token replacement for the token input and mask fields (ai_integration_eca, MR) — https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26 [25] Add missing OpenAI TTS voices to api_defaults.yml (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3522771 [26] Add test coverage for Whisper (speechToText) and TTS (textToSpeech) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894 [27] Support batched embeddings (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568659 [28] Moderation endpoint has hardcoded model (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418 [29] "Selected API Key not working" after update to AI 1.3.0 (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3579428 [30] Add reasoning output (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3574226 [31] Change the way that system prompt is being added in 1.2.x (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3573133 [32] Add AI usage checkboxes to ai_provider_openai issue template (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568620 [33] Add AI 2.x / Search AI compatible OpenAI embedding provider (2.x branch) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3564086 [34] Support File Handling (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3548526 [35] Remove OpenAI SDK dependency and extend it from Drupal AI core module (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531260 [36] Handle pdf files in chat (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3529603 [37] Version of modules can go out of alignment (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519419 [38] Abstract token usage support (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519302 [39] Usage stats in streaming chat need conversion to an array. (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486224 [40] Log flagged prompts (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3507407 [41] Moderation log overview (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3516044 [42] Streaming chat responses gives Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools() (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3515885 [43] Allow skipping of moderations for some embeddings (not all) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3510601 [44] Bug: Key validation and Rate Limit tester are bypassing host variable; causing Auth errors or Network errors, etc. (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3496412 [45] Selecting "o1" model on automator produces error: "Unsupported parameter: 'max_tokens' is not supported" (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3508042 [46] Error with OpenAI o1 models (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489201 [47] Moderation reads confidence scores instead of the actual violation flags, so almost every category shows as flagged (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161 [48] Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat() (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601 [49] Add support for fast mode (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211 [50] Unable to do GPT 5.4 calls (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729 [51] How to enable type: web_search in Chat? (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3613029 [52] Automated Drupal 12 compatibility fixes for ai_provider_openai 1.3.x-dev (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3606117 [53] Create a skill to setup the provider (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3605812 [54] Automated Drupal 12 compatibility fixes for ai_provider_openai 1.2.x-dev (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3594765 [55] Opt-in to cache embedding lookups (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3552523 [56] Turn off moderation if omni-moderation-latest is not available (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3534855 [57] Add gpt-image-1 image to image capabilitites (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531289 [58] Warning appears while using automators for image type field "Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() " (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3504849 [59] Add event triggering for streaming (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489846 [60] Improve documentation on setting key (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486487 [61] Enable Drupal AI Module Functionality for Environments Behind a Proxy (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3481664 [62] Reasoning models return an empty message, and model-specific parameters are rejected by Converse (ai_provider_aws_bedrock, Issue) — https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/work_items/3594715 [63] Issue #3594715: Read the first content block that carries text, and route model-specific parameters to additionalModelRequestFields (ai_provider_aws_bedrock, MR) — https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19 [64] Structured output: an invalid JSON schema is silently accepted instead of throwing an exception (ai_provider_azure, Issue) — https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013 [65] Add batch embeddings (ai_vdb_provider_postgres, Issue) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3610509 [66] indexItems() throws TypeError on array_diff() because $skipDeleteItemIds is never declared (ai_vdb_provider_postgres, Issue) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3614504 [67] Issue #3614504: Fix TypeError in indexItems() from undeclared $skipDeleteItemIds property (ai_vdb_provider_postgres, MR) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33 [68] feat: #3610509 Add batch embeddings (ai_vdb_provider_postgres, MR) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30 [69] Dependency Dashboard (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591657 [70] [>=11.4.7] Broaden our check to determine if a recipe is being applied, use 11.4.7's new `RecipeRunner::isApplying()` (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591985 [71] RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591996 [72] [PP-1] Add logic + docs to not ship Component config entities with Drupal extensions (modules, themes, recipes), and how to do it anyway when there's a good enough reason (pre-populated component trees in Canvas Pages or ContentTemplates) (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3565712 [73] Allow Canvas Patterns+PageRegions+ContentTemplates provided by Recipes to NOT specify component versions (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3571366 [74] Canvas AI: Add Playwright coverage for the chat Tools dropdown (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591951 [75] Config in language other than site default language causes an exception (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591850 [76] Canvas AI: Proper chat history management in the backend (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592032 [77] Canvas AI: Prevent the selected Tool from changing during a chat turn (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591952 [78] Site template installation fails with Canvas 1.11 (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592052 [79] Canvas AI: Unify error handling in `canvas_component_agent`'s tools (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592021 [80] Remove Canvas-only text format guidelines from non-Canvas forms (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592061 [81] <iframe sandbox> + `Window.postMessage()` to add defense in depth (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3527515 [82] Canvas AI: Update orchestrator for content-only / page-variant chrome (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592060 [83] Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592030 [84] Normalize agent-written prop values against their schema and report values the component source cannot store (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592002 [85] Canvas AI: Unify error handling in the dev page builder agent's tools (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591958 [86] Add component label to settings form (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591934 [87] The error message about "collapsed input syntax" could be more helpful (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3558318 [88] Check for the recipe runner (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1554 [89] refactor(Conflict resolution): #3591785 Strengthen AutoSaveManager to generate stable hashes (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1356 [90] Add docs/recipes.md: how Canvas Component versions interact with recipes... (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1634 [91] Draft: Resolve "Non HTML-validating raw template fails to add Canvas slot wrapper comments" (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1638 [92] #3592031: Added padding to the container to fix visual gaps. (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1600 [93] Draft: 🤖 feat: #3592009 Add locale-aware date/time formatting filters (Twig) and functions (JS) (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1632 [94] 🤖 bug: #3591850 Use entity langcode, not site default, as default translation (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1400 [95] Draft: Resolve "Support mapping multi-valued list fields to array-type component props" (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1630 [96] Improve the Brand Kit Fonts UI (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1520 [97] #3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1604 [98] chore(deps): update dependency cspell to v10 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1273 [99] chore(deps): update dependency @eslint/compat to v2 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1306 [100] chore(deps): update dependency @chromatic-com/storybook to v5 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1305 [101] fix(deps): update npm minor and patch (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1238 [102] chore(deps): update mysql docker tag to v8.3 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1272 [103] bug: #3592061 Remove Canvas-only text format guidelines from non-Canvas forms (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1637 [104] Iframe sandbox: drop allow-same-origin, add postMessage defense in depth (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1610 [105] Draft: task(AI): #3592060 Canvas AI: Update orchestrator for content-only / page-variant chrome (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1636 [106] task(AI): #3592030 Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1624 [107] 3592021: Add AiResponseValidator::formatErrors() to render accumulated tool... (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1626 [108] issue #3591965: add condition to just load lib on required theme (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1529 [109] test(AI): #3591951 Canvas AI: Add Playwright coverage for the chat Tools dropdown (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1607 [110] feat(AI): #3592032 Canvas AI: Keep the agent history across chat turns in the backend (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1635 [111] Unify error handling in all tools used by the dev page builder agent. (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1584 [112] chore: #3592007 Allow developer and test tooling to locate the Drupal root when the module checkout lives outside the Drupal tree (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1573 [113] Sync brand kit colors through canvas.brand-kit.json in the Canvas CLI (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1586 [114] fix(Data model): #3558318 The error message about "collapsed input syntax" could be more helpful (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1633 [115] Issue #3592055 by f0ns: Hide auto-created folders whose Components are all disabled (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1628 [116] fix(CLI Tool): #3592059 Improve Canvas CLI compatibility with Canvas <1.11 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1631 [117] CCC Pre-rc1 QA: Context items settings testing (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586496 [118] Disabling "Create new revisions by default" still creates a new revision (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586515 [119] Update context consumer listing configuration summary column (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586514 [120] Change Add AI Context Item button to Add Context Item (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586513 [121] Accessibility Testing: CCC overview dashboard (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586434 [122] CCC Pre-rc1 QA: Test context item creation and workflow (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586491 [123] CCC Pre-rc1 QA: Context scope testing (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586492 [124] Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections (ai_ckeditor, Issue) — https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848 [125] Issue #3615848: Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections. (ai_ckeditor, MR) — https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16 [126] How often is the ecosystem data updated? (ai_dashboard, Issue) — https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221 [127] Issue #3585221: Rebuild the ecosystem (ai_dashboard, MR) — https://git.drupalcode.org/project/ai_dashboard/-/merge_requests/34 [128] Target element lookup fails for select widgets (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828 [129] Handle empty "Set widget field value" gracefully (with helpful message) (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578827 [130] Bug when trying to "Remove action"? (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578826 [131] Configured actions are invisible in the "Manage form display" summary row (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578830 [132] TranslatableMarkup returned as string causes fatal error (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578829 [133] Add action summary to manage fields ui (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/35 [134] #3578829: Cast return value to string in getDescription() (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/34 [135] Add action summary to manage fields ui (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/32 [136] #3578829: Cast return value to string in getDescription() (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/33 [137] HTML markup is corrupted/not preserved when translating formatted text fields (text_with_summary, text_long) (ai_translate, Issue) — https://git.drupalcode.org/project/ai_translate/-/work_items/3585525 [138] Regression: Layout Builder content is no longer translated (ai_translate, Issue) — https://git.drupalcode.org/project/ai_translate/-/work_items/3585553 [139] Resolve "Regression: Layout Builder content is no longer translated" (ai_translate, MR) — https://git.drupalcode.org/project/ai_translate/-/merge_requests/49 [140] Text-to-speech audio is truncated to the first ~5,000 characters because MP3 chunks are joined by raw byte concatenation (elevenlabs, Issue) — https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581 [141] Issue #3597581 by banoodle: Fix audio truncated to the first chunk when text... (elevenlabs, MR) — https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5 [142] Bare editor/playground page leaks <head-placeholder>, pushing every stylesheet into <body> so UI themes (Drafter, Minimal) and skins never apply (flowdrop, Issue) — https://git.drupalcode.org/project/flowdrop/-/work_items/3592452 --- MODULE SUMMARIES --- <h3>AI (Artificial Intelligence)</h3> <p>The module saw significant activity around Field Widget Actions and bug fixes. The <a href="https://git.drupalcode.org/project/ai/-/work_items/3586655">Add interactive refinement capability to Field Widget Actions 2.x</a> port was merged via two separate merge requests (<a href="https://git.drupalcode.org/project/ai/-/merge_requests/2002">!2002</a> and <a href="https://git.drupalcode.org/project/ai/-/merge_requests/1966">!1966</a>), bringing the interactive refinement feature from 1.x into the 2.x branch. This required adapting the implementation to 2.x's plugin collection architecture, including a new <code>AutomatorFieldWidgetActionTrait</code> and <code>AutomatorRefineAction</code> class. Related work continues on <a href="https://git.drupalcode.org/project/ai/-/work_items/3586657">Allow Field Widget Actions to target a specific automator 2.x</a>, which adds an optional <code>$specificAutomatorId</code> parameter to <code>AiAutomatorEntityModifier::saveEntity()</code> for filtering execution to a specific automator.</p> <p>Several critical bugs were identified during the period. The <a href="https://git.drupalcode.org/project/ai/-/work_items/3586730">textToImage() base64-decodes the image payload twice</a> regression in <code>OpenAiBasedProviderClientBase::textToImage()</code> silently corrupts generated images by double-decoding the payload. The <a href="https://git.drupalcode.org/project/ai/-/work_items/3586729">SetupVdbServer and SetupVdbIndex config actions aren't idempotent</a> issue prevents recipe re-application due to missing entity existence checks in the config actions' <code>apply()</code> methods. Additionally, <a href="https://git.drupalcode.org/project/ai/-/work_items/3586624">Streaming chat forces plain-text hostname filtering</a> was reported, where <code>StreamChatMessageIteratorTrait</code> breaks URLs inside HTML attributes by forcing plain-text mode in <code>HostnameFilter</code>.</p> <h4>How can I help on this project?</h4> <p>Review the open merge request <a href="https://git.drupalcode.org/project/ai/-/merge_requests/2001">!2001</a> for the textToImage double-decode fix and help verify the fix with test coverage. Pick up the unassigned <a href="https://git.drupalcode.org/project/ai/-/work_items/3586732">Content Entity Seeder JSON Schema bug</a>, which involves fixing <code>ContextDefinitionNormalizer</code> wrapping issues. Assist with the assigned <a href="https://git.drupalcode.org/project/ai/-/work_items/3586733">AI API Explorer permission security issue</a> by reviewing the proposed permission restriction.</p> Contributors: Abhisek Mazumdar (<a href="https://www.drupal.org/u/abhisekmazumdar">abhisekmazumdar</a>), Artem Dmitriiev (<a href="https://www.drupal.org/u/a.dmitriiev">a.dmitriiev</a>), Dezső Biczó (<a href="https://www.drupal.org/u/mxr576">mxr576</a>), Juan Correa (<a href="https://www.drupal.org/u/jucs7">jucs7</a>), Petar Bašić (<a href="https://www.drupal.org/u/petar_basic">petar_basic</a>), Sven Decabooter (<a href="https://www.drupal.org/u/svendecabooter">svendecabooter</a>) --- <h3>AI Integration - ECA</h3> <p>The AI Integration - ECA module saw significant activity during the reporting period, with three issues closed and a major refactor merged. The module reached a stable 1.0.0 release milestone, making it compatible with Drupal CMS recipes that require minimum-stability: stable. This addresses the <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025">Stable 1.0.0 release for use in Drupal CMS recipes</a> request and enables broader adoption in packaged distributions.</p> <p>A token handling refactor was completed with the merge of <a href="https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26">MR <a href="https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26">!26</a></a>, which addresses <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027">Replace token references by token replacements</a>. The change standardizes token input and mask fields to use token replacement rather than token references throughout the module, following the pattern established in earlier work. The commit touched 1355 lines of code.</p> <p>Support for media entities in vision models was also completed during this period, closing <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289">Support media entities for vision models</a>. This enhancement enables media field files to be sent to AI models with prompts for categorization.</p> <h4>How can I help on this project?</h4> <p>With the stable release shipped, contributors could test integration with Drupal CMS recipes and report any issues encountered. Testing the new media entity support with various vision models and file types would help validate functionality across different AI providers.</p> <p>Contributors: Jürgen Haas (<a href="https://www.drupal.org/u/jurgenhaas">jurgenhaas</a>) [1]</p> --- <h3>OpenAI Provider</h3> <p>The OpenAI Provider saw significant housekeeping activity over the past 24 hours, with 24 issues closed as part of ongoing maintenance. Several closed issues relate to feature completions from earlier development cycles, including <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3522771">Add missing OpenAI TTS voices to api_defaults.yml</a>, <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568659">Support batched embeddings</a>, and <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3574226">Add reasoning output</a> for o1-series models. Bug fixes included resolving <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3579428">"Selected API Key not working" after update to AI 1.3.0</a> and <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3515885">streaming chat responses giving undefined getTools() method errors</a>.</p> <p>Active development continues on several fronts. A merge request for <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894">test coverage for Whisper (speechToText) and TTS (textToSpeech)</a> is under review, adding 43 new tests for audio operation paths. Two critical bugs remain open: <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161">moderation reading confidence scores instead of violation flags</a>, causing false positives, and <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418">hardcoded model in moderationEndpoint</a> that should respect configuration. Feature requests include <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211">fast mode support</a> for GPT models and addressing <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729">GPT 5.4 call failures</a> related to reasoning_effort parameters.</p> <h4>How can I help on this project?</h4> <p>Review the open merge request for <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894">Whisper and TTS test coverage</a>. Investigate and patch the <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161">moderation confidence score bug</a>, which needs to read flagged boolean values instead of score floats. Refactor the <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418">hardcoded moderationEndpoint model</a> to respect operation-type configuration with fallback logic.</p> --- <h3>AWS Bedrock Provider</h3> <p>A bug was identified affecting reasoning models in the AWS Bedrock Provider. The issue concerns how the module reads responses from Amazon Bedrock's Converse API. Currently, <code>chat()</code> reads only the first content block, assuming it contains text. However, reasoning models return their reasoning in a separate content block that arrives first, causing the module to return an empty <code>ChatMessage</code> when it encounters <code>reasoningContent</code> instead of <code>text</code>. Additionally, model-specific parameters passed through the <code>config</code> array are rejected by the Converse API, which expects them in the <code>additionalModelRequestFields</code> structure.</p> <p>A merge request addressing both issues is currently open: <a href="https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19">Issue <a href="https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/work_items/3594715">#3594715</a>: Read the first content block that carries text, and route model-specific parameters to additionalModelRequestFields</a>. The proposed solution scans content blocks to find the first one containing text and routes model-specific parameters to the correct API field.</p> <h4>How can I help on this project?</h4> <p>Review and test the open <a href="https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19">merge request addressing reasoning model responses</a>, particularly with different Bedrock models that use reasoning content blocks. Verify the parameter routing logic works correctly with various model-specific configurations.</p> <p>Contributors: Besart Bytyqi (<a href="https://www.drupal.org/u/besart">besart</a>)</p> --- <h3>Microsoft Azure AI</h3> <p>No merge requests or commits were pushed during the reporting period. However, a schema validation issue was reported: <a href="https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013">Structured output: an invalid JSON schema is silently accepted instead of throwing an exception</a>. The bug describes how malformed JSON schemas—where required properties are not defined in the properties object—are passed through to the Azure API without local validation, preventing developers from catching configuration errors early.</p> <p>The issue remains open with no related merge requests or assigned maintainer. This represents a data integrity concern for sites using structured output features, as invalid schemas may produce unpredictable results from the Azure provider without clear error messages.</p> <h4>How can I help on this project?</h4> <ul> <li>Pick up <a href="https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013">the JSON schema validation bug</a> and implement client-side schema validation before forwarding requests to Azure.</li> <li>Test the module's structured output functionality with edge cases and document any additional validation gaps.</li> </ul> --- <h3>Postgres VDB Provider</h3> <p>Development activity this period focused on batch embedding support and a critical bugfix. Work continued on <a href="https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3610509">Add batch embeddings</a>, which aligns this provider with the AI module's new batch embeddings capability for improved indexing performance. The related merge request remains under review, introducing the ability to embed and store multiple chunks per request rather than processing one at a time.</p> <p>A new issue was opened reporting a TypeError in the indexItems() method: <a href="https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3614504">indexItems() throws TypeError on array_diff() because $skipDeleteItemIds is never declared</a>. The problem occurs because the $skipDeleteItemIds property is referenced on line 685 but never declared or initialized in PostgresProvider or its parent class, causing PHP to treat it as null and triggering an array_diff() type error during indexing operations. A fix has been prepared addressing the undeclared property.</p> <h4>How can I help on this project?</h4> <p>Review <a href="https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30">MR <a href="https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30">!30</a></a> for batch embeddings support, testing the indexing performance improvements with multiple chunks. Review <a href="https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33">MR <a href="https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33">!33</a></a> which fixes the TypeError in indexItems() by properly declaring the $skipDeleteItemIds property.</p> Contributors: Sergiu Nagailic (<a href="https://www.drupal.org/u/Nikro">Nikro</a>), Zeeshaan khann (<a href="https://www.drupal.org/u/zeeshan_khan">zeeshan_khan</a>) --- <h3>Drupal Canvas</h3> <p>Activity over the past 24 hours centered on the Canvas AI chat tooling, compatibility improvements, and several open merge requests targeting error handling and user experience refinements.</p> <p>The fix to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591952">prevent the selected Tool from changing during a chat turn</a> was merged, locking the tool selection for the entire turn in both the UI and backend. A critical bug causing recipe application to fail with a synthetic service error—<a href="https://git.drupalcode.org/project/canvas/-/work_items/3591996">RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it</a>—is under review and appears related to Canvas 1.11 site template installation failures reported in <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592052">Site template installation fails with Canvas 1.11</a>.</p> <p>Several AI-focused merge requests remain in progress, including unified error handling for component agent tools in <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592021">Canvas AI: Unify error handling in canvas_component_agent's tools</a> and backend chat history management in <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592032">Canvas AI: Proper chat history management in the backend</a>. Work also continues on removing obsolete region-based instructions from the page and template builder agents in <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592030">Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form</a>.</p> <p>Other notable open requests include improved Canvas CLI compatibility with Canvas versions prior to 1.11, iframe sandbox hardening via <code>postMessage</code> defense in depth, and a fix ensuring Canvas-only text format guidelines no longer appear on non-Canvas forms.</p> <h4>How can I help on this project?</h4> <ul> <li>Review the open merge request for <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591996">RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it</a>, which is blocking recipe application and site template installs.</li> <li>Test and review <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592061">Remove Canvas-only text format guidelines from non-Canvas forms</a>, which has an open MR ready for feedback.</li> </ul> <p>Contributors: **** (<a href="https://www.drupal.org/u/project_19391_bot_cb443e60680723bba08a5351bf595251">project_19391_bot_cb443e60680723bba08a5351bf595251</a>), Adam G-H (<a href="https://www.drupal.org/u/phenaproxima">phenaproxima</a>), Ahmad Khalil (<a href="https://www.drupal.org/u/ahmad-khalil-imagex">ahmad-khalil-imagex</a>) [1], AKHIL BABU (<a href="https://www.drupal.org/u/AkhilBabu">AkhilBabu</a>), Bálint Kléri (<a href="https://www.drupal.org/u/balintbrews">balintbrews</a>), Chandan Singh (<a href="https://www.drupal.org/u/chandu7929">chandu7929</a>), <a href="https://www.drupal.org/u/dpi">dpi</a>, Feliksas Mazeikis (<a href="https://www.drupal.org/u/f.mazeikis">f.mazeikis</a>), Fons Vandamme (<a href="https://www.drupal.org/u/f0ns">f0ns</a>), Georgi Ivanov (<a href="https://www.drupal.org/u/givanov">givanov</a>), John Locke (<a href="https://www.drupal.org/u/freelock">freelock</a>), Jordan Graham (<a href="https://www.drupal.org/u/COBadger">COBadger</a>), julio retkwa (<a href="https://www.drupal.org/u/julioretkwa">julioretkwa</a>), Konstantine Kirkitadze (<a href="https://www.drupal.org/u/kotekirkitadze">kotekirkitadze</a>), Lauri Timmanee (<a href="https://www.drupal.org/u/lauriii">lauriii</a>), Matt Hand (<a href="https://www.drupal.org/u/matthand">matthand</a>), Narendra Singh Rathore (<a href="https://www.drupal.org/u/NarendraR">NarendraR</a>), Ted Bowman (<a href="https://www.drupal.org/u/tedbow">tedbow</a>), utkarsh_33 (<a href="https://www.drupal.org/u/Utkarsh_33">Utkarsh_33</a>), Wim Leers (<a href="https://www.drupal.org/u/wimleers">wimleers</a>), <a href="https://www.drupal.org/u/wotnak">wotnak</a></p> --- <h3>Context Control Center (CCC)</h3> <p>No code changes were merged or committed during this period. Activity focused on pre-RC1 quality assurance and issue triaging.</p> <p>A new bug was filed: <a href="https://git.drupalcode.org/project/ai_context/-/work_items/3586515">Disabling "Create new revisions by default" still creates a new revision</a>. The revision setting at <code>/admin/config/ai/context/settings/items</code> is not being respected when existing AI Context Items are edited and saved. This appears to have been discovered during structured QA testing.</p> <p>Multiple QA test plans covering context items, scopes, and settings remain in progress. One workflow test issue was closed. Two UX refinement tasks were opened: renaming the "Add AI Context Item" button and updating the context consumer listing to show subscription details in the configuration summary column.</p> <p>Accessibility testing of the CCC overview dashboard is also underway, focusing on WCAG compliance checks including keyboard navigation and DOM structure.</p> <h4>How can I help on this project?</h4> <ul> <li>Investigate and patch the revision setting bug in <a href="https://git.drupalcode.org/project/ai_context/-/work_items/3586515">Disabling "Create new revisions by default" still creates a new revision</a>, checking where the entity save logic ignores the configured default.</li> <li>Pick up QA testing on <a href="https://git.drupalcode.org/project/ai_context/-/work_items/3586496">CCC Pre-rc1 QA: Context items settings testing</a> or <a href="https://git.drupalcode.org/project/ai_context/-/work_items/3586492">CCC Pre-rc1 QA: Context scope testing</a> following the provided test scripts.</li> </ul> --- <h3>AI CKEditor</h3> <p>Activity this period focused on improving entity context handling for AI-powered text generation. Work continued on <a href="https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848">Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections</a>, which addresses a bug where AI Context scoping fails in two scenarios: when editing unsaved entities (where the entity ID is empty and validateEntityContext() returns NULL), and when using the Site Sections module (where the entity being edited differs from the page's main entity). The issue has an open merge request that refactors how entity_context metadata is constructed and transmitted, ensuring bundle and path information are available even without a saved entity ID. No merges occurred during this period, though the MR remains under consideration for the AI Initiative Sprint.</p> <h4>How can I help on this project?</h4> <p>Review <a href="https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16">MR <a href="https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16">!16</a></a> to validate the entity_context refactor handles unsaved entities and Site Sections correctly. Test the changes against AI Context scoping rules to confirm context items are properly matched during text generation.</p> <p>Contributors: Kristen Pol (<a href="https://www.drupal.org/u/kepol">kepol</a>)</p> --- <h3>AI Dashboard</h3> <p>The <a href="https://git.drupalcode.org/project/ai_dashboard/-/merge_requests/34">Issue <a href="https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221">#3585221</a>: Rebuild the ecosystem</a> merge request was merged during this period, refreshing the ecosystem data to include 234 modules and 38 recipes as of September 10, 2026. The update also included a fix to the crawler to handle raw socket timeouts with retry logic. This addresses the question raised in <a href="https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221">How often is the ecosystem data updated?</a>, with the ecosystem data now current as of the merge date.</p> <h4>How can I help on this project?</h4> <p>The AI Dashboard currently has limited open issues. Developers interested in contributing could review the ecosystem crawler implementation to identify additional edge cases or timeouts that may need handling, or explore the refreshed ecosystem data to verify accuracy and completeness of the 234 modules and 38 recipes now tracked.</p> <p>Contributors: Rob Loach (<a href="https://www.drupal.org/u/robloach">robloach</a>) [1]</p> --- <h3>Field Widget Actions</h3> <p>Two fixes were merged during the reporting period. A fatal error caused by a type mismatch in <code>FieldWidgetActionBase::getDescription()</code> was resolved by casting the TranslatableMarkup return value to string, fixing <a href="https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578829">TranslatableMarkup returned as string causes fatal error</a>. The change updates the method to properly handle the plugin definition's description attribute.</p> <p>A usability enhancement was also shipped: <a href="https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578830">Configured actions are invisible in the "Manage form display" summary row</a> was closed after merging work that adds action summaries directly to the Manage form display interface. This makes it easier to identify which fields have actions configured without opening individual widget settings.</p> <p>Three issues remain open. <a href="https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828">Target element lookup fails for select widgets</a> tracks a bug in <code>FieldWidgetActionBase::getTargetElement()</code> where the hardcoded 'value' path fails for select, radios, checkboxes, and other multiple-value widgets. Two additional issues address error handling when removing actions and validating empty widget field values.</p> <h4>How can I help on this project?</h4> <p>Fix the target element lookup logic in <code>FieldWidgetActionBase::getTargetElement()</code> to support select and other multiple-value widgets as described in <a href="https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828">Target element lookup fails for select widgets</a>. Investigate the "Remove action" bug reported in <a href="https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578826">Bug when trying to "Remove action"?</a> to identify the root cause.</p> Contributors: Artem Dmitriiev (<a href="https://www.drupal.org/u/a.dmitriiev">a.dmitriiev</a>), Joonas Meriläinen (<a href="https://www.drupal.org/u/mErilainen">mErilainen</a>), Nik LePage (<a href="https://www.drupal.org/u/niklp">niklp</a>) --- <h3>AI translate</h3> <p>A regression affecting Layout Builder translations was identified and addressed during the reporting period. Since the changes introduced in the 1.3.2 and 1.4.1 security releases, Layout Builder content has silently failed to translate. The issue stems from <code>TextExtractor::shouldExtract()</code> performing an <code>access('view')</code> check on every field, while core's <code>LayoutSectionItemList::defaultAccess()</code> returns <code>AccessResult::forbidden()</code> unconditionally for all users, including user 1. This causes the <code>layout_builder__layout</code> field to be dropped before <code>LbFieldExtractor</code> can process it, preventing inline block content from being translated with both <code>layout_builder_at</code> and <code>layout_builder_st</code>.</p> <p>A merge request was opened to <a href="https://git.drupalcode.org/project/ai_translate/-/merge_requests/49">resolve the Layout Builder regression</a>, introducing <code>ContainerFieldTextExtractor</code> to handle the layout field appropriately.</p> <p>Separately, a long-standing issue remains open regarding <a href="https://git.drupalcode.org/project/ai_translate/-/work_items/3585525">HTML markup corruption in formatted text fields</a>, where AI models strip or alter HTML tags in <code>text_with_summary</code> and <code>text_long</code> fields during translation.</p> <h4>How can I help on this project?</h4> <p>Review and test the <a href="https://git.drupalcode.org/project/ai_translate/-/merge_requests/49">Layout Builder regression fix</a> to validate it resolves the access check issue. Investigate the <a href="https://git.drupalcode.org/project/ai_translate/-/work_items/3585525">HTML markup preservation problem</a> by proposing prompt engineering strategies or post-processing filters to maintain tag integrity.</p> Contributors: Sven Decabooter (<a href="https://www.drupal.org/u/svendecabooter">svendecabooter</a>), Carlos Busto Capeáns (<a href="https://www.drupal.org/u/cbccharlie">cbccharlie</a>) --- <h3>ElevenLabs</h3> <p>A critical bug affecting audio generation was identified during the reporting period. The <a href="https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581">Text-to-speech audio is truncated to the first ~5,000 characters because MP3 chunks are joined by raw byte concatenation</a> issue documents a problem in ElevenlabsProvider::textToSpeech() where audio files are being cut short despite text being within the 10,000 character API limit. The function splits longer text into 5,000-character chunks and requests separate MP3 files from the ElevenLabs API, but the current implementation naively concatenates the binary data of complete MP3 files using string append operations. This approach fails because each MP3 response includes its own headers and metadata, resulting in only the first chunk playing correctly. A fix has been proposed via MR <a href="https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5">!5</a> with 282 lines of changes addressing the concatenation logic.</p> <h4>How can I help on this project?</h4> <p>Review <a href="https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5">MR <a href="https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5">!5</a></a> which fixes the MP3 chunk concatenation bug. Test the fix with text exceeding 5,000 characters to verify audio plays completely and transitions between chunks are seamless.</p> <p>Contributors: Anne Bonham (<a href="https://www.drupal.org/u/annebonham">annebonham</a>)</p> --- <h3>FlowDrop</h3> <p>The bare editor/playground page renderer bug that caused stylesheets to leak into the document body has been fixed. The <code>BarePageRenderer</code> in flowdrop_ui_components was emitting a literal <code><head-placeholder></code> tag; because the bare page attaches no <code>html_head</code> elements, core's <code>HtmlResponseAttachmentsProcessor</code> never replaced the placeholder. The HTML parser then closed <code><head></code> prematurely, forcing all <code><link></code> tags into <code><body></code> and preventing UI themes (Drafter, Minimal) and skins from applying. The <a href="https://git.drupalcode.org/project/flowdrop/-/work_items/3592452">bare page head placeholder fix</a> was committed.</p> <p>A chat tool loop issue was also resolved: tool-heavy conversation turns no longer wedge when hitting the memory cap. The JavaScript library was bumped to version 2.8.0, and new site-wide switches were added for the Command Console and AI Assistant surfaces, with the defunct Sync to Cloud toggle removed. The 2.6.0-alpha1 release was re-stamped to include the chat tool loop fix, and changelog entries were prepared covering the chat tool loop, editor surface switches, and the bare page head fix.</p> <h4>How can I help on this project?</h4> <p>Review the recent commits for the <a href="https://git.drupalcode.org/project/flowdrop/-/work_items/3592452">bare page head placeholder fix</a> to ensure edge cases around attachment processing are covered. Test the new Command Console and AI Assistant surface switches across different site configurations to confirm the settings persist correctly.</p> <p>Contributors: Shibin Das [8]</p> ======================================================================== ## [16] factcheck:tldr:developer ======================================================================== Layer-1 corrections: 3 Judge flags: 3 ======================================================================== ## [17] summariseModule:ai:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI (Artificial Intelligence) (machine name: ai) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Abhisek Mazumdar (abhisekmazumdar), Artem Dmitriiev (a.dmitriiev), Dezső Biczó (mxr576), Juan Correa (jucs7), Petar Bašić (petar_basic), Sven Decabooter (svendecabooter)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI (Artificial Intelligence)</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [LLM: Metatag generates and stores tags that were never configured, and keeps only one of them](https://git.drupalcode.org/project/ai/-/work_items/3586723) State: opened | Updated: 2026-09-11 | Author: Ahmad Khader (Ahmad-Khader) | Assigned: unassigned | Labels: category::bug, priority::normal, state::rtbc Description: ## Summary Two defects in the Metatag automator, both independent of the decoding regression fixed in #3586713 and both invisible until that one is fixed, because until then nothing was stored at all. 1. **Every tag is offered to the model, not the configured ones.** `extraAdvancedFormFields()` renders a textarea for every tag Metatag knows about, so the saved configuration carries a key for all of them. `generate()` builds its "available tags" list from those keys rather than from the ones holding a subprompt. The form is explicit that an empty one means *"Keep empty to not generate anythin… ### [Allow Field Widget Actions to target a specific automator 2.x](https://git.drupalcode.org/project/ai/-/work_items/3586657) State: opened | Updated: 2026-09-11 | Author: Artem Dmitriiev (a.dmitriiev) | Assigned: unassigned | Labels: 2026Sprint16, 2026Sprint17, AI Initiative Sprint, ai::inside Description: Port changes from issue https://git.drupalcode.org/project/ai/-/work_items/3571915 to 2.x and 2.0.x branches ### [Review AI issue process and tagging based on recent usage and update docs](https://git.drupalcode.org/project/ai/-/work_items/3550025) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: unassigned | Labels: category::task, documentation, priority::normal, state::needsWork Description: >>> [!note] Migrated issue Reported by: [kristen pol](https://www.drupal.org/user/8389) >>> Problem/Motivation Review the docs and figure out what needs changing based on how we've been using tags/etc: https://project.pages.drupalcode.org/ai/1.1.x/contribute/issue_guidelines Proposed resolution Update docs as necessary Remaining tasks Review, tweak, review, commit > Related issue: [Issue #3540998](https://www.drupal.org/node/3540998) ### [Create AI assistant API session only after the first message](https://git.drupalcode.org/project/ai/-/work_items/3554797) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, AI Innovation, aiAssistantsApi, category::bug Description: >>> [!note] Migrated issue Reported by: [omarlopesino](https://www.drupal.org/user/2853043) Related to !978 >>> Problem/Motivation Currently, when the chatbot is added in the site for anonymous users, a new session is started in order to create the thread. This is necessary, but it can be optimized. Creating a session invalidates reverse proxy caches like Varnish or any kind of CDN, decreasing the site's performance. The session may be only created after the first interaction with the chatbot. Steps to reproduce (required for bugs, but not feature requests) Create an assistant with AI a… ### [Add interactive refinement capability to Field Widget Actions 2.x](https://git.drupalcode.org/project/ai/-/work_items/3586655) State: closed | Updated: 2026-09-11 | Author: Artem Dmitriiev (a.dmitriiev) | Assigned: unassigned | Labels: 2026Sprint16, 2026Sprint18, ai::inside, aiAutomators Description: This issue is for cherry-picking / porting the changes from related issue to 2.x. As in 2.x the way ai_automator works was slightly change, it is not a straightforward cherry-pick. MR to port/cherry-pick https://git.drupalcode.org/project/ai/-/merge_requests/1629 ### [Automator FWA buttons give no user feedback when values don't visibly change](https://git.drupalcode.org/project/ai/-/work_items/3585987) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: unassigned | Labels: 2026Sprint16, 2026Sprint17, AI Initiative Sprint, On hold Description: >>> [!note] Migrated issue Reported by: [petar_basic](https://www.drupal.org/user/3626336) >>> [Tracker] Update Summary: [One-line status update for stakeholders] Short Description: [One-line issue summary for stakeholders] Check-in Date: MM/DD/YYYY Metadata is used by the AI Tracker. Docs and additional fields here. [/Tracker] Problem/Motivation When an editor clicks a Field Widget Action (FWA) button backed by an ai_automator, the AJAX call runs, the throbber appears and disappears, but there is no user feedback confirming that the automator actually ran. This is especially visible on th… ### [Restrict access to the AI API Explorer permission](https://git.drupalcode.org/project/ai/-/work_items/3586733) State: opened | Updated: 2026-09-11 | Author: Marcus Johansson (Marcus_Johansson) | Assigned: Prabhavathi Vanipenta (prabha1997) | Labels: Novice, Quick Win, Release blocker, aiApiExplorer Description: The AI API Explorer is a pure developer tool, but its `access ai prompt` permission is not marked as security-sensitive. Anyone holding it can invoke AI providers directly, and with provider function calling being added it becomes a way to exercise function calls outside the access checks of the features that normally wrap them. That is a meaningful attack surface if the permission is ever granted to a non-developer role. Proposed change in `modules/ai_api_explorer/ai_api_explorer.permissions.yml`: ```yaml access ai prompt: title: 'Permit access to using the AI API Explorer.' description… ### [SetupVdbServer and SetupVdbIndex config actions aren't idempotent, recipe re-apply fails](https://git.drupalcode.org/project/ai/-/work_items/3586729) State: opened | Updated: 2026-09-10 | Author: Abhisek Mazumdar (abhisekmazumdar) | Assigned: Ahmad Khalil (ahmad-khalil-imagex) | Labels: AI Initiative Sprint, ai::inside, category::bug, priority::normal Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !2000 [fix: #3586729 Add existence checks to `SetupVdbServer` and `SetupVdbIndex` config actions](https://git.drupalcode.org/project/ai/-/merge_requests/2000) | State: opened | Author: Abhisek Mazumdar (abhisekmazumdar), 441 diff lines Description: ## Problem/Motivation `SetupVdbServer::apply()` and `SetupVdbIndex::apply()` both call `->create($value)->save()` unconditionally, with no check for whether the target entity already exists: ```php // SetupVdbServer.php, line 111-113 $this->entityTypeManager->getStorage('search_api_server') ->create($value) ->save(); ``` ```php // SetupVdbIndex.php, line 53-55 $this->entityTypeManager->getStorage('search_api_index') ->create($value) ->save(); ``` Any recipe that uses these config actions (e.g. `ai_recipe_content_search_vector`) can't be applied twice on the same site. Re-applying i… ### [Adding Disable Option for AI Automators](https://git.drupalcode.org/project/ai/-/work_items/3508800) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: aiAutomators, category::feature, priority, priority::normal Description: >>> [!note] Migrated issue Reported by: [vakulrai](https://www.drupal.org/user/3524052) Related to !490 >>> Problem/Motivation While working on the AI automator's i felt there is no way we could disable the automator configured on a field , rather than we have to go to edit the field and remove the automator checkbox in order to remove from the automator chain. the drawback i see is admin's have to reconfigure the same set of changes again if they later on enable the same features. Is it a good idea to add this feature and make sure admins can enable and disable the automator's by clickin… ### [Bug — the "Content Entity Seeder" tool generates a JSON Schema with one extra array level and breaks when creating/editing content](https://git.drupalcode.org/project/ai/-/work_items/3586732) State: opened | Updated: 2026-09-10 | Author: Bruno Carvalho (brunocarvalho) | Assigned: unassigned | Labels: category::bug, priority::normal, state::accepted Description: ### **Environment** | **Item** | **Value** | |----------|-----------| | Drupal | 11.4.6 | | PHP | 8.4 | | `drupal/ai` | 1.4.8 | | `drupal/ai_agents` | 1.3.5 (the same `execute()` is present on the `1.4.x` branch / `1.4.0-beta1`) | | Provider / model | OpenAI `gpt-5.4` (via `ai_provider_openai` 1.2.5) | | Tool | `ai_agent:content_entity_seeder` → `Drupal\ai_agents\Plugin\AiFunctionCall\ContentEntitySeeder` | --- ### **Description** `ContextDefinitionNormalizer` wraps the `ContentEntitySeeder`'s `entity_array` parameter in one extra array level (`multiple` + `ComplexToolItems` applied togeth… ### [textToImage() base64-decodes the image payload twice](https://git.drupalcode.org/project/ai/-/work_items/3586730) State: opened | Updated: 2026-09-10 | Author: Sven Decabooter (svendecabooter) | Assigned: unassigned | Labels: aiCoreModule, category::bug, priority::major, state::accepted Description: ## Summary `OpenAiBasedProviderClientBase::textToImage()` calls `base64_decode()` a second time on data it has already decoded, so the `ImageFile` it returns holds corrupt bytes instead of the generated image. The failure is silent: the mime-type guard immediately above inspects the *correct* bytes and passes, and the corrupt bytes are handed out afterwards, so callers get a successful response and write unusable files. The same line also throws away the mime type it just validated in favour of a hardcoded `image/png`. This is a regression. Before commit bc860ac9 ("Let providers select allow… ### [Regression in ai_translate: Layout Builder content is no longer translated](https://git.drupalcode.org/project/ai/-/work_items/3586731) State: opened | Updated: 2026-09-10 | Author: Sven Decabooter (svendecabooter) | Assigned: unassigned | Labels: No Estimation Available, category::bug, priority::critical, translation Description: This issue has been reported in the standalone version of ai_translate module, but the ai_translate module that ships as a submodule within this module is also affected by this: https://git.drupalcode.org/project/ai_translate/-/work_items/3585553 Once a 1.3.x fix for this issue exists in standalone ai_translate, it should probably be backported here as well. Feel free to review / help move the original issue forward, so a fix can be quickly applied here as well. ### [Streaming chat forces plain-text hostname filtering, corrupting URLs inside HTML attributes (e.g., href="")](https://git.drupalcode.org/project/ai/-/work_items/3586624) State: opened | Updated: 2026-09-10 | Author: Ron Raney (rraney) | Assigned: unassigned | Labels: 2026Sprint17, AI Initiative Sprint, On hold, ai::inside Description: **Problem/Motivation** `StreamChatMessageIteratorTrait::getHostnameFilterService()` forces `HostnameFilter::setPlainTextMode(TRUE)` whenever code isn't running inside a PHP Fiber (which is the normal case for typical request-based streaming consumers, e.g. AI CKEditor's Reformat/Modify tools). With plain-text mode forced on, `HostnameFilter::filterText()` skips the HTML-aware filtering branch and instead runs `filterUrlsInPlainText()` — a raw regex substring replace across the entire response text, with no awareness of HTML structure. When AI-processed content contains an `` tag pointing to … --- MERGE REQUESTS (24h) --- ### [Issue #3586657: Allow Field Widget Actions to target a specific automator 2.x](https://git.drupalcode.org/project/ai/-/merge_requests/1999) State: opened | Author: Juan Correa (jucs7) | Branch: 3586657-allow-field-widget-actions-target-specific-automator-2x | 185 diff lines Description: ## Description **Changes:** - `AiAutomatorEntityModifier::saveEntity()` gains an optional `$specificAutomatorId` parameter. When set, processing is filtered to the matching automator config entity ID before any processor runs. - Unlike 1.x, 2.x's `entityHasAutomators()` already returns automators keyed by config entity ID and sorts them with a key-preserving `uasort()`, so the filter here operates directly on that keyed array instead of 1.x's `array_filter()` placed before a key-destroying `usort()`. - Added `testSaveEntityFiltersBySpecificAutomatorId()` and `testSaveEntityReturnsNullForUnma… ### [Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x"](https://git.drupalcode.org/project/ai/-/merge_requests/2002) State: merged 2026-09-11 | Author: Artem Dmitriiev (a.dmitriiev) | Branch: cherry-pick-3d89b17d | 2979 diff lines Description: ## Description ## Testing instructions 1. 2. ## Checklist - [ ] I have linked the related issue in the MR title or description - [ ] I have performed a self-review of my own code - [ ] I have added or updated tests, or explained in the description why this change is not covered by tests - [ ] I have updated documentation for any new or changed functionality - [ ] I have written testing instructions and verified them locally - [ ] I have noted any required post-merge steps (config imports, cache rebuilds, manual changes) - [ ] This MR contains no breaking API or hook changes, or the… ### [feat: #3585987 by petar.basic: Add status message after AI Field Widget Action automator runs.](https://git.drupalcode.org/project/ai/-/merge_requests/1540) State: opened | Author: Petar Bašić (petar_basic) | Branch: 3585987-add-feedback-message | 79 diff lines Description: Closes #3585987 ### [Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x"](https://git.drupalcode.org/project/ai/-/merge_requests/1966) State: merged 2026-09-11 | Author: Juan Correa (jucs7) | Branch: 3586655-add-interactive-refinement-capability-to-field-widget-actions-2x | 2967 diff lines Description: ## Summary Ports the interactive refinement capability for Field Widget Actions Automation from `1.x` (issue #3586468, MR !1629) to `2.x`. As noted in the issue, this was **not** a straight cherry-pick: in `2.x` an `ai_automator` config entity holds an `automator_types` **plugin collection** (each entry its own configured plugin instance) instead of `1.x`'s flat `rule` + `plugin_config`. The port re-implements the same feature against that architecture. ### Changes - **New `Traits/AutomatorFieldWidgetActionTrait.php`** , extracted from `AutomatorBaseAction` so both it and the new `Automato… ### [Draft: fix: Fix AI Settings form to display AI Platforms and Inference Providers](https://git.drupalcode.org/project/ai/-/merge_requests/1901) State: opened | Author: Dezső Biczó (mxr576) | Branch: 3586631-ai-platform-providers-does-not-appear-in-ai-default-settings | 60 diff lines Description: ## Description Closes #3586631 ## TODO - [ ] Test coverage ## Testing instructions 1. 2. ## Checklist - [ ] I have linked the related issue in the MR title or description - [ ] I have performed a self-review of my own code - [ ] I have added or updated tests, or explained in the description why this change is not covered by tests - [ ] I have updated documentation for any new or changed functionality - [ ] I have written testing instructions and verified them locally - [ ] I have noted any required post-merge steps (config imports, cache rebuilds, manual changes) - [ ] This MR co… ### [bug: #3586730 textToImage() base64-decodes the image payload twice](https://git.drupalcode.org/project/ai/-/merge_requests/2001) State: opened | Author: Sven Decabooter (svendecabooter) | Branch: 3586730-texttoimage-base64-decodes-the-image-payload-twice | 189 diff lines Description: ## Description Closes #3586730. `OpenAiBasedProviderClientBase::textToImage()` decoded the `b64_json` payload twice, so the `ImageFile` it returned held bytes that are not an image. The second decode keeps roughly a quarter of the payload, and because the mime-type guard directly above it inspects the *correctly* decoded bytes and passes, the call still reports success — the corruption only surfaces when the file is written and opened. The change is one line in `src/Base/OpenAiBasedProviderClientBase.php`: ```php - $images[] = new ImageFile(base64_decode($image_content), 'image/png', 'gene… --- COMMITS (24h) --- (none) ======================================================================== ## [18] summariseModule:ai_integration_eca:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI Integration - ECA (machine name: ai_integration_eca) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Jürgen Haas (jurgenhaas) [1]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI Integration - ECA</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Support media entities for vision models](https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: Marcus Johansson (Marcus_Johansson) | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai::inside Description: >>> [!note] Migrated issue Reported by: [anaconda777](https://www.drupal.org/user/3595801) >>> Problem/Motivation Hi, Currently the ECA Chat task has a text area for prompt. (I am currently using it for categorizing user generated text and it works great) For vision models, there is a need to send media items (like jpg, png etc. files) to the model for categorizing them. So if a content contains a Media field, the file of it should be send to a Model with a prompt. (This by the way works already in Chat Generation Explorer where you can type a prompt and under that is a file upload. ) So a… ### [Stable 1.0.0 release for use in Drupal CMS recipes](https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025) State: closed | Updated: 2026-09-10 | Author: Rajab Natshah (RajabNatshah) | Assigned: unassigned | Labels: category::task, priority::normal, state::fixed Description: Thank you so much Jürgen and Marcus for this module. 934 sites report using it. We added it to our products, in the AI base recipe. We would love to ship it in Drupal CMS recipes. Recipes install with `minimum-stability: stable`, and 1.0.x is rc only, so the install stops: ``` Root composer.json requires drupal/ai_integration_eca ~1, found drupal/ai_integration_eca[dev-1.0.x, 1.0.0-alpha2, ..., 1.0.x-dev (alias of dev-1.0.x)] but it does not match your minimum-stability. ``` You are already at rc5, so a stable 1.0.0 tag would open the door for everyone packaging recipes and products. Happy… ### [Replace token references by roken replacements](https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027) State: closed | Updated: 2026-09-10 | Author: Jürgen Haas (jurgenhaas) | Assigned: unassigned | Labels: category::bug, code, priority::major, state::fixed Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !26 [fix: #3585027 Use token replacement for the token input and mask fields](https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26) | State: merged 2026-09-10 | Author: Jürgen Haas (jurgenhaas), 1355 diff lines Description: In project/ai_integration_eca#3534289 we've changed the file field from not using token references but token replacement. That should also be done everywhere else in the module. --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- - [204a3af8](https://git.drupalcode.org/project/ai_integration_eca/-/commit/204a3af86dbffeb4c36e38d83fd1dc21efab21df) fix: #3585027 Use token replacement for the token input and mask fields — Jürgen Haas (2026-09-10T13:07:04.000+00:00) ======================================================================== ## [19] summariseModule:ai_provider_openai:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: OpenAI Provider (machine name: ai_provider_openai) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. Do not name any individual people in this section. No one was active enough this period to list. Format your response as an HTML fragment. Start with <h3>OpenAI Provider</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Add missing OpenAI TTS voices to api_defaults.yml](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3522771) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [dan_metille](https://www.drupal.org/user/574844) Related to !32 >>> Thank you for your great work on this module! I noticed that the current list of OpenAI text-to-speech voices in api_defaults.yml only includes 6 options, while the official OpenAI API now supports 11 voices. I've created a patch (attached) that adds the missing voices: ash, ballad, coral, sage, and verse. This ensures all currently available OpenAI TTS voices can be used with the module. ### [Add test coverage for Whisper (speechToText) and TTS (textToSpeech)](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [alekas](https://www.drupal.org/user/726268) Related to !83 >>> Problem/Motivation The speechToText() (Whisper) and textToSpeech() (TTS) operation overrides in OpenAiProvider currently have zero test coverage. The existing 44 tests only cover Chat, streaming, tool calls, the File API and moderation. Any regression in the audio operations (multipart payload serialization, response parsing, error mapping) would go unnoticed. Proposed resolution Add 43 tests using a real openai-php/client over a Guzzle MockHandler transport — no live API calls, f… ### [Support batched embeddings](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568659) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: 2026Sprint14, 2026Sprint15, AI Initiative Sprint, ai inside Description: >>> [!note] Migrated issue Reported by: [paulsheldrake](https://www.drupal.org/user/1350686) Related to !59 >>> Update the module to support batches for embeddings so they can be inserted into VDBs faster Works with these patches: AI Milvus ### [Moderation endpoint has hardcoded model](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [a.dmitriiev](https://www.drupal.org/user/3235287) >>> Problem/Motivation The method moderationEndpoint is using hardcoded 'omni-moderation-latest' model. This method was there before the moderation operation type was introduced in AI Core module. Now it should of course use the model, that is selected for this operation type. Of course there should be a check that the selected provider is OpenAI. The hardcoded model should be used as a fallback. ### ["Selected API Key not working" after update to AI 1.3.0](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3579428) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, category::bug, code, priority::major Description: >>> [!note] Migrated issue Reported by: [aschiwi](https://www.drupal.org/user/107667) Related to !70 >>> Problem/Motivation This might be a problem caused by the AI module instead of OpenAI Provider but I'm encountering it on an OpenAI Provider config page: I was adding the AI module to an existing site and could not add an OpenAI API key because at admin/config/ai/providers/openai it says this when selecting the valid key "The selected API key is not working. Please double-check the correct API key was entered and that it has credit(s) available.". I tried several things and was confuse… ### [Add reasoning output](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3574226) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !68 >>> Problem/Motivation With the issue https://git.drupalcode.org/project/ai/-/work_items/3546573 we are adding the reasoning summaries as part of the output, so you can see what the model did reason. OpenAI will work as test bed to see that this is working. Steps to reproduce Proposed resolution Introduce reasoning in none-streaming mode. Introduce reasoning in streaming mode if its available while streaming. Remaining tasks User interface changes API changes Data model changes >… ### [Change the way that system prompt is being added in 1.2.x](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3573133) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: Novice, category::task, code, priority::normal Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !65 >>> Problem/Motivation The way historically we set a system prompt was due to the LLM being the main focus at the time, but it was obvious later that the system prompt should be set on the ChatInput instead. This is documented here: https://project.pages.drupalcode.org/ai/1.2.x/developers/develop_third_party_module/#setting-chat-system-messages We should switch over so the OpenAI Provider does this. Proposed resolution Switch from $this->chatSystemRole to fetching it from the i… ### [Add AI usage checkboxes to ai_provider_openai issue template](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568620) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, miscellaneous, priority, priority::major Description: >>> [!note] Migrated issue Reported by: [kristen pol](https://www.drupal.org/user/8389) >>> [Tracker] Update Summary: [One-line status update for stakeholders] Check-in Date: MM/DD/YYYY Additional Collaborators: Metadata is used by the AI Tracker. Docs and additional fields here. [/Tracker] Problem/Motivation Parent: https://git.drupalcode.org/project/ai_initiative/-/work_items/3568605 Issues must declare their AI usage, or they violate DA policies as noted in: Abuse of the Contribution Credit system Proposed resolution Update the issue template with checkboxes from the parent issue. Targ… ### [Add AI 2.x / Search AI compatible OpenAI embedding provider (2.x branch)](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3564086) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [mbatterton](https://www.drupal.org/user/1574576) Related to !57 !58 >>> AI framework 2.x introduces new provider and capability interfaces, which are not compatible with the existing 1.2.x OpenAI provider. This issue introduces a new 2.0.x development branch that: - Implements EmbeddingProviderInterface for AI 2.x - Is compatible with search_ai 2.x - Supports OpenAI embedding models - Leaves 1.2.x fully intact for AI 1.x users This is an embeddings-only provider to support AI Search and vector indexing use cases. AI 2.x separates capabilities expli… ### [Support File Handling](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3548526) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, AI Product Development, category::feature, code Description: >>> [!note] Migrated issue Reported by: [b_sharpe](https://www.drupal.org/user/2512258) Related to !79 !53 >>> Problem/Motivation See https://platform.openai.com/docs/api-reference/files/create Once https://git.drupalcode.org/project/ai/-/work_items/3546954 lands, we can add the ability to upload files and use them for various purposes. Eventually this could be added automatically in some cases (chats with history, fine-tuning). > Related issue: [Issue #3546954](https://www.drupal.org/node/3546954) ### [Remove OpenAI SDK dependency and extend it from Drupal AI core module](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531260) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [gxleano](https://www.drupal.org/user/3591999) Related to !44 >>> Problem/Motivation Currently, OpenAI provider is implemented using the OpenAI SDK directly, with only minor variations. This approach has led to significant code duplication (reused by other providers), especially evident when implementing new features like function calling. In such cases, nearly identical logic had to be added in multiple places, making the codebase harder to maintain, test, and evolve. To address this, we already have an AiProviderClientBase class as a common founda… ### [Handle pdf files in chat](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3529603) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [bircher](https://www.drupal.org/user/1344166) Related to !42 >>> Problem/Motivation Openai can handle pdfs in some models: https://platform.openai.com/docs/guides/pdf-files?api-mode=chat Steps to reproduce Attach a pdf to a chat. Proposed resolution Handle PDFs attached to chat. Remaining tasks adjust the messages when pdfs are attached to chat. User interface changes none API changes none Data model changes none > Related issue: [Issue #3529601](https://www.drupal.org/node/3529601) ### [Version of modules can go out of alignment](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519419) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [mrdalesmith](https://www.drupal.org/user/2612656) Related to !29 >>> Problem/Motivation This is a duplicate of https://www.drupal.org/project/ai_agents/issues/3518506#comment-16067498 to resolve the issue for the main provider in use by people whilst a more generic solution is come up with. Description copied from that issue. In theory you will be able to update the modules in the following order, because of the dependency tree: * AI 1.1.x * AI Agents 1.1.x * Some provider still on 1.0.x This leads to the provider not having any function calling av… ### [Abstract token usage support](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519302) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [aspilicious](https://www.drupal.org/user/172527) Related to !30 >>> https://www.drupal.org/project/ai/issues/3517618 > Related issue: [Issue #3524435](https://www.drupal.org/node/3524435) > Related issue: [Issue #3526936](https://www.drupal.org/node/3526936) > Related issue: [Issue #3526978](https://www.drupal.org/node/3526978) > Related issue: [Issue #3517618](https://www.drupal.org/node/3517618) ### [Usage stats in streaming chat need conversion to an array.](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486224) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [shaundychko](https://www.drupal.org/user/475828) Related to !3 >>> Problem/Motivation \Drupal\ai\OperationType\Chat\StreamedChatMessage::__construct expects the third parameter to be an array. However, when a streaming chat is created with usage included, such as with the following example configuration:       /** @var \Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider $provider */      $provider->setConfiguration([        'stream_options' => [ … ### [Log flagged prompts](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3507407) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::rtbc Description: >>> [!note] Migrated issue Reported by: [mindbet](https://www.drupal.org/user/1703106) Related to !21 >>> Problem/Motivation When loading content in bulk to a vector database, certain paragraphs may be flagged by moderation. The interface does not show exactly what content caused the issue so that it can be addressed. Proposed resolution Add code similar to this to OpenAiProvider.php: (lines 626-628) if (!empty($response['results'][0]['flagged'])) { $moderation_alert = 'This prompt was flagged by the moderation model: ' . $prompt; \Drupal::logger('ai_provider_openai')->… ### [Moderation log overview](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3516044) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [jonas139](https://www.drupal.org/user/2873401) Related to !26 >>> Problem/Motivation When using moderation an error will be thrown when a prompt is flagged. In issue https://www.drupal.org/project/ai_provider_openai/issues/3507407 there's already a logging mechanism but the error is still thrown so it's not very user friendly. It seems better to not throw an error but log the flagged prompt in a separate overview that can be used to track down the related entity (if there is any) so a content manager can review the flagged prompt and rewrite the pi… ### [Streaming chat responses gives Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools()](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3515885) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [cellear](https://www.drupal.org/user/599892) Related to !27 >>> Problem/Motivation From fresh install of Drupal CMS (1.1x-dev) from the Chat Explorer (admin/config/ai/explorers/chat_generator) with Streamed checked, the following error appears: The website encountered an unexpected error. Try again later.Error: Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools() in Drupal\ai_api_explorer\Plugin\AiApiExplorer\ChatGenerator->getResponse() (line 306 of modules/contrib/ai/modules/ai_api_explorer/src/Plugin/AiApiE… ### [Allow skipping of moderations for some embeddings (not all)](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3510601) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, needs forward port, priority::normal Description: >>> [!note] Migrated issue Reported by: [wouters_f](https://www.drupal.org/user/721548) Related to !23 >>> Problem/Motivation Allow the skipping of moderations for some requests, not all. > Related issue: [Issue #3510599](https://www.drupal.org/node/3510599) ### [Bug: Key validation and Rate Limit tester are bypassing host variable; causing Auth errors or Network errors, etc.](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3496412) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [jan kellermann](https://www.drupal.org/user/371731) Related to !43 >>> Problem/Motivation We try to use an openAI compatible API. In OpenAiProvider the config-item `host` is supported - see code. But in OpenAiConfigForm the OpenAi-Client is loaded without custom host - see code so a choise of models is not possible. Proposed resolution Add UI for Host Use \OpenAI::factory() like here. > Related issue: [Issue #3531134](https://www.drupal.org/node/3531134) ### [Selecting "o1" model on automator produces error: "Unsupported parameter: 'max_tokens' is not supported"](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3508042) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, needs backport to 1.1.x, priority::normal Description: >>> [!note] Migrated issue Reported by: [alfthecat](https://www.drupal.org/user/495982) Related to !28 !22 >>> Problem/Motivation Hi, I'm using an automator on a list field and when I switch it to use the "o1" model instead of my previously selected "40o" model under advanced settings, I get the following errors in the log and the field does not get automated/interpolated. "Error invoking model response: Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead." "A general error happened why trying to interpolate, message Error invoking mod… ### [Error with OpenAI o1 models](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489201) State: closed | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::closed Description: >>> [!note] Migrated issue Reported by: [merilainen](https://www.drupal.org/user/304437) >>> Problem/Motivation I tested the new Open o1 models in AI Chat Explorer, got an error: The following error occurred while processing your request:Error invoking model response: Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.I tied also without max_tokens value but got the same error. ### [Moderation reads confidence scores instead of the actual violation flags, so almost every category shows as flagged](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [rduterte](https://www.drupal.org/user/3809153) >>> Problem/Motivation OpenAI's moderation endpoint returns two different things per category: a confidence score (a number, always present, usually small and non-zero even for harmless text) and a flag (a true or false stating whether that category actually tripped the moderation threshold). This provider's own moderation method reads the confidence scores and stores them as if they were the flags. Since PHP treats any non-zero number as true, and the scores are almost never exactly zero, every categor… ### [Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat()](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [kaszarobert](https://www.drupal.org/user/3578328) >>> Problem/Motivation Using 1.3.0-beta2 throws an error for all calls with the latest stable ai module 1.4.8: Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat() (line 319 of /app/web/modules/contrib/ai_provider_openai/src/Plugin/AiProvider/OpenAiProvider.php)#0 [internal function]: Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat(Object(Drupal\ai\OperationType\Chat\ChatInput), 'gpt-5.6-terra', Array)#… ### [Add support for fast mode](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [scott_euser](https://www.drupal.org/user/3267594) >>> Problem/Motivation A good chunk of the models support fast mode (seems roughly 2x the cost, for 2.5x the response speed) Steps to reproduce See https://developers.openai.com/api/docs/pricing?latest-pricing=fast Proposed resolution Consider how to support; maybe the providers list needs to return them as separate items e.g. gpt 5.4 mini, gpt 5.4 mini (fast mode) as it does not seem desireable to have an all-or-nothing configuration on the OpenAI provider settings made Remaining tasks Decide/agree … ### [Unable to do GPT 5.4 calls](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [jesmaster](https://www.drupal.org/user/1494544) >>> I have my model set to 5.4-nano and it was working fine until this recent update. Now I get "Function tools with reasoning_effort are not supported for gpt-5.4-nano in /v1/chat/completions." If switch to 5.6 it works fine. Most likely related to https://www.drupal.org/project/ai_provider_openai/issues/3613982 ### [How to enable type: web_search in Chat?](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3613029) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::support, documentation, priority::normal, state::postponed Description: >>> [!note] Migrated issue Reported by: [anybody](https://www.drupal.org/user/291091) >>> Postponed on https://git.drupalcode.org/project/ai/-/work_items/3586616 Problem/Motivation In a project we're using Drupal AI for research, including web sources. We're listing different sources to search for in our prompt and we're using OpenAI Provider. I was a bit curious if AI really searches the web for the results (and I'm still) and I found https://developers.openai.com/api/docs/guides/tools-web-search Can someone tell if type: web_search is required to be enabled for AI looking into current r… ### [Automated Drupal 12 compatibility fixes for ai_provider_openai 1.3.x-dev](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3606117) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: Drupal 12 compatibility, ProjectUpdateBotD12, category::task, code Description: >>> [!note] Migrated issue Reported by: [project update bot](https://www.drupal.org/user/3644742) Related to !77 >>> Problem/Motivation Hello project maintainers, This is an automated issue to help make this module compatible with Drupal 12. Changes will periodically be added to this issue that remove deprecated API uses. To stop further changes from being posted, change the status to anything other than Active, Needs review, Needs work or Reviewed and tested by the community. Alternatively, you can remove the "ProjectUpdateBotD12" tag from the issue to stop the bot from posting updates. … ### [Create a skill to setup the provider](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3605812) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: AI Initiative Sprint, AI Innovation, category::feature, code Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !82 >>> Description We want an agent skill (under ai_provider_openai/.agents/skills/) that sets up the OpenAI provider from an API key that already exists on the system as an environment variable. The skill must work without ever exposing the key's value: it checks only that the variable is set, then wires it in through the Key module's environment key provider so the secret never lands in configuration. This skill does not use the Tool API — it is driven entirely through drush … ### [Automated Drupal 12 compatibility fixes for ai_provider_openai 1.2.x-dev](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3594765) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: Drupal 12 compatibility, ProjectUpdateBotD12, category::task, code Description: >>> [!note] Migrated issue Reported by: [project update bot](https://www.drupal.org/user/3644742) Related to !75 >>> Problem/Motivation Hello project maintainers, This is an automated issue to help make this module compatible with Drupal 12. Changes will periodically be added to this issue that remove deprecated API uses. To stop further changes from being posted, change the status to anything other than Active, Needs review, Needs work or Reviewed and tested by the community. Alternatively, you can remove the "ProjectUpdateBotD12" tag from the issue to stop the bot from posting updates. … ### [Opt-in to cache embedding lookups](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3552523) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::postponed Description: >>> [!note] Migrated issue Reported by: [scott_euser](https://www.drupal.org/user/3267594) Related to !55 >>> Problem/Motivation https://git.drupalcode.org/project/ai/-/work_items/3552522 provides trait method to cache embeddings lookups. Steps to reproduce Embeddings method does not use it Proposed resolution Use the new caching option Remaining tasks Decide 2.0.x branch vs method_exists check User interface changes N/A API changes N/A Data model changes N/A ### [Turn off moderation if omni-moderation-latest is not available](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3534855) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !49 >>> Problem/Motivation When working with a company account in OpenAI, there is the possibility to only allow certain models. This means that you can turn off the moderation models. This is very bad practice since the accounts can get banned, but we have one reported instance of such an account. We need the OpenAI model to still work in these instances and people might not know that we do the opinionated thing of turning on moderation by default. We need to inform when this happens… ### [Add gpt-image-1 image to image capabilitites](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531289) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !45 >>> Problem/Motivation We have added an image to image operation type in https://git.drupalcode.org/project/ai/-/work_items/3531212, meaning that we will be able to take an image and defined parameters and use them and generate an image. OpenAI gpt-image-1 can do this. We should implement it. Steps to reproduce Proposed resolution Implement the ImageToImage operation type with gpt-image-1. Remaining tasks User interface changes API changes Data model changes ### [Warning appears while using automators for image type field "Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() "](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3504849) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::postponed Description: >>> [!note] Migrated issue Reported by: [vakulrai](https://www.drupal.org/user/3524052) >>> Problem/Motivation When i enabled image automator to generate image and selected "default Image to text model" the node save was triggering warnings and not able to generate and save the image in the field. Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() (line 404 of /var/www/html/docroot/modules/contrib/ai_provider_openai/src/Plugin/AiProvider/OpenAiProvider.php) Steps to reproduce Enable and configure the Openai module add a… ### [Add event triggering for streaming](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489846) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, code, priority::normal, state::needsReview Description: >>> [!note] Migrated issue Reported by: [marcus_johansson](https://www.drupal.org/user/385947) Related to !48 >>> Problem/Motivation Add the methods that makes logging possilbe of responses on streaming events from here: https://www.drupal.org/project/ai/issues/3489690 Steps to reproduce Proposed resolution Remaining tasks User interface changes API changes Data model changes ### [Improve documentation on setting key](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486487) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::task, documentation, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [jvdkolk](https://www.drupal.org/user/562054) >>> The section 'Post-Installation' states: "Set up an API key using the Key module with your OpenAI credentials." Nevertheless, on admin/config/system/keys/add there is a section 'Type settings' - Key type (4 types) - Provider settings - etc. What to set there? I also had to clear the cache before my key showed up on admin/config/ai/providers/openai Proposed text "Set up an API key using the Key module with your OpenAI credentials. Under 'Key type', choose 'Authentication'. Under 'Key provider' choose '… ### [Enable Drupal AI Module Functionality for Environments Behind a Proxy](https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3481664) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::feature, code, priority::normal, proxy Description: >>> [!note] Migrated issue Reported by: [olegiv](https://www.drupal.org/user/3268706) >>> Problem/Motivation The Drupal AI module currently does not support environments where requests must be routed through an HTTP/HTTPS proxy. Many organizations use proxies to secure and manage external communications, especially in development and corporate environments. Without proper proxy support, external API calls made by the module may fail, rendering the module unusable in these environments. Steps to reproduce 1. Set up Drupal in a local or corporate environment where internet access is restrict… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [20] summariseModule:ai_provider_aws_bedrock:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AWS Bedrock Provider (machine name: ai_provider_aws_bedrock) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Besart Bytyqi (besart)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AWS Bedrock Provider</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Reasoning models return an empty message, and model-specific parameters are rejected by Converse](https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/work_items/3594715) State: opened | Updated: 2026-09-10 | Author: Besart Bytyqi (besart) | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::needsReview Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !19 [Issue #3594715: Read the first content block that carries text, and route model-specific parameters to additionalModelRequestFields](https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19) | State: opened | Author: Besart Bytyqi (besart), 50 diff lines Description: ## Issue summary ### Problem 1: the response is empty for any reasoning model `chat()` reads the answer from content block zero: ```php $message = new ChatMessage($response['output']['message']['role'], $response['output']['message']['content'][0]['text'] ?? ""); ``` Reasoning models put their reasoning in its own content block, and it arrives first. Block zero holds `reasoningContent`, not `text`, so the `?? ""` fires and the provider returns an empty `ChatMessage`. Nothing errors, so the caller sees a successful call that produced no text. That `?? ""` came from #3517795, where block ze… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [21] summariseModule:ai_provider_azure:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Microsoft Azure AI (machine name: ai_provider_azure) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. Do not name any individual people in this section. No one was active enough this period to list. Format your response as an HTML fragment. Start with <h3>Microsoft Azure AI</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Structured output: an invalid JSON schema is silently accepted instead of throwing an exception](https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013) State: opened | Updated: 2026-09-10 | Author: Marcus Johansson (Marcus_Johansson) | Assigned: unassigned | Labels: category::bug, code, priority::normal, v1.2.x-dev Description: ## Summary Structured output calls do not fail when the supplied JSON schema is itself invalid. A schema whose `required` array lists a key that is not defined in `properties` is forwarded to the Azure API and the call succeeds, so the caller never learns that the schema is broken. The schema below should not be usable — `description` is listed as required but never defined, while `caption` is defined but never required: ```json { "type": "object", "properties": { "alt": { "type": "string", "description": "Concise alt text for accessibility (max 125 characters)" }, … --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [22] summariseModule:ai_vdb_provider_postgres:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Postgres VDB Provider (machine name: ai_vdb_provider_postgres) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Sergiu Nagailic (Nikro), Zeeshaan khann (zeeshan_khan)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>Postgres VDB Provider</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Add batch embeddings](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3610509) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: 2026Sprint14, 2026Sprint16, AI Initiative Sprint, ai inside Description: >>> [!note] Migrated issue Reported by: [nikro](https://www.drupal.org/user/576416) Related to !30 >>> Problem/Motivation The AI module is introducing batch embeddings as a first-class capability, so indexing can embed and store multiple chunks per request instead of one at a time. This meaningfully reduces indexing time. See the parent work item #3568648: Batch embeddings for improved indexing performance. Vector database providers are being aligned with this capability so they participate in the batched indexing path. The Milvus provider has already been updated in ai_vdb_provider_milvu… ### [indexItems() throws TypeError on array_diff() because $skipDeleteItemIds is never declared](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3614504) State: opened | Updated: 2026-09-10 | Author: Zeeshaan khann (zeeshan_khan) | Assigned: unassigned | Labels: Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !33 [Issue #3614504: Fix TypeError in indexItems() from undeclared $skipDeleteItemIds property](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33) | State: opened | Author: Zeeshaan khann (zeeshan_khan), 18 diff lines Description: ### Problem/Motivation Indexing any content through a Search API index backed by this provider fails with: TypeError: array_diff(): Argument #2 must be of type array, null given in PostgresProvider.php on line 685 This happens on every call to PostgresProvider::indexItems(), because the property $this->skipDeleteItemIds is read but never declared or initialized anywhere in the class (or in its parent, AiVdbProviderClientBase). PHP treats the undefined property access as null, which array_diff() rejects as an invalid second argument. ### Steps to reproduce Step 1: Set up a Search API serve… --- MERGE REQUESTS (24h) --- ### [feat: #3610509 Add batch embeddings](https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30) State: opened | Author: Sergiu Nagailic (Nikro) | Branch: 3610509-add-batch-embeddings | 656 diff lines Description: By: nikro Closes #3610509 --- COMMITS (24h) --- (none) ======================================================================== ## [23] summariseModule:canvas:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Drupal Canvas (machine name: canvas) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: **** (project_19391_bot_cb443e60680723bba08a5351bf595251), Adam G-H (phenaproxima), Ahmad Khalil (ahmad-khalil-imagex) [1], AKHIL BABU (AkhilBabu), Bálint Kléri (balintbrews), Chandan Singh (chandu7929), dpi, Feliksas Mazeikis (f.mazeikis), Fons Vandamme (f0ns), Georgi Ivanov (givanov), John Locke (freelock), Jordan Graham (COBadger), julio retkwa (julioretkwa), Konstantine Kirkitadze (kotekirkitadze), Lauri Timmanee (lauriii), Matt Hand (matthand), Narendra Singh Rathore (NarendraR), Ted Bowman (tedbow), utkarsh_33 (Utkarsh_33), Wim Leers (wimleers), wotnak". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>Drupal Canvas</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Dependency Dashboard](https://git.drupalcode.org/project/canvas/-/work_items/3591657) State: opened | Updated: 2026-09-11 | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Assigned: unassigned | Labels: projectManagement Description: This issue lists Renovate updates and detected dependencies. Read the [Dependency Dashboard](https://docs.renovatebot.com/key-concepts/dashboard/) docs to learn more. ## Repository Problems Renovate tried to run on this repository, but found these problems. - ⚠️ WARN: Failed to set branch status - ⚠️ WARN: No github.com token has been configured. Skipping release notes retrieval - ⚠️ WARN: Rate limit exceeded for api.github.com, as no hostRules set for this host. Please set a GITHUB_COM_TOKEN - ⚠️ WARN: No tool releases found. ## Deprecations / Replacements > ⚠️ **Warning** > The foll… ### [[>=11.4.7] Broaden our check to determine if a recipe is being applied, use 11.4.7's new `RecipeRunner::isApplying()`](https://git.drupalcode.org/project/canvas/-/work_items/3591985) State: opened | Updated: 2026-09-11 | Author: Adam G-H (phenaproxima) | Assigned: unassigned | Labels: Needs issue summary update, Needs upstream bugfix, Recipes initiative, category::task Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1554 [Check for the recipe runner](https://git.drupalcode.org/project/canvas/-/merge_requests/1554) | State: opened | Author: Adam G-H (phenaproxima), 86 diff lines Description: Right now, we use a pretty narrow hack to determine if a recipe is being applied: https://git.drupalcode.org/project/canvas/-/blob/1.x/src/ComponentSource/ComponentSourceManager.php?ref_type=heads#L100 As the comment notes, the "config is syncing" flag is inconsistent about recipe application. There's no core API to differentiate a normal config sync from a config sync happening during a recipe. Core _will_ add that API in https://www.drupal.org/project/drupal/issues/3613607, but it's not done yet. So until then, we need this hack. But the hack isn't broad enough; we need to know if a recipe… ### [RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it](https://git.drupalcode.org/project/canvas/-/work_items/3591996) State: opened | Updated: 2026-09-11 | Author: Jakub Piasecki (zaporylie) | Assigned: unassigned | Labels: category::bug, priority::critical, state::needsReview Description: ## Overview Applying a site template recipe fails with: In ContainerBuilder.php line 1104: You have requested a synthetic service ("kernel"). The DIC does not know how to construct this service. `RecipeSubscriber::onApply()` calls `PageVariantMigration::migrateDefaultTheme()`, which calls `createFromPageRegions()`, which installs `canvas_page_template_component` when it is missing: ``` // src/PageVariantMigration.php:131 if (!\Drupal::moduleHandler()… ### [[PP-1] Add logic + docs to not ship Component config entities with Drupal extensions (modules, themes, recipes), and how to do it anyway when there's a good enough reason (pre-populated component trees in Canvas Pages or ContentTemplates)](https://git.drupalcode.org/project/canvas/-/work_items/3565712) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: Adam G-H (phenaproxima), Wim Leers (wimleers) | Labels: DX (Developer Experience), Recipes initiative, category::task, configManagement Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1634 [Add docs/recipes.md: how Canvas Component versions interact with recipes...](https://git.drupalcode.org/project/canvas/-/merge_requests/1634) | State: opened | Author: Wim Leers (wimleers), 158 diff lines Description: > [!NOTE] > Issue summary needs updating. Incorporate #3564135+'s different focus. >>> [!note] Migrated issue Reported by: [wim leers](https://www.drupal.org/user/99777) >>> Overview Ever since Component versions were introduced (#3523841: Versioned Component config entities (SDC, JS: prop_field_definitions, block: default_setting, all: slots for fallback) + component instances refer to versions ⇒ less data to store per XB field row), it became implicitly unsupported to ship an exported Component config entity with some Drupal extension in its config/install or config/optional direct… ### [Allow Canvas Patterns+PageRegions+ContentTemplates provided by Recipes to NOT specify component versions](https://git.drupalcode.org/project/canvas/-/work_items/3571366) State: opened | Updated: 2026-09-11 | Author: drupalbot | Assigned: Matt Glaman (mglaman) | Labels: AI-accelerated, Chicago2026, DX (Developer Experience), Drupal CMS dependency critical Description: >>> [!note] Migrated issue Reported by: [penyaskito](https://www.drupal.org/user/959536) Related to !811 >>> Overview In https://git.drupalcode.org/project/canvas/-/work_items/3567413 we added the infra for updating component instances to the latest (active) version if there is no data impact. That is pretty defensive per nature, so we ensure no data loss. But there are some cases where e.g. contrib should be able to impact that. e.g. In \Drupal\canvas\Hook\ShapeMatchingHooks::mediaLibraryStorablePropShapeAlter we implement hook_canvas_storable_prop_shape_alter for media_library, and for… ### [Canvas AI: Add Playwright coverage for the chat Tools dropdown](https://git.drupalcode.org/project/canvas/-/work_items/3591951) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: Konstantine Kirkitadze (kotekirkitadze) | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1607 [test(AI): #3591951 Canvas AI: Add Playwright coverage for the chat Tools dropdown](https://git.drupalcode.org/project/canvas/-/merge_requests/1607) | State: opened | Author: Konstantine Kirkitadze (kotekirkitadze), 202 diff lines Description: ## Summary https://git.drupalcode.org/project/canvas/-/work_items/3591856 added a Tools dropdown to the dev chat UI. Install `canvas_dev_ai`, open the chat, and the dropdown is there. Nothing covers it in a browser. Add Playwright coverage. ## Current behavior No browser coverage exists for the dropdown. The tool list, the selection, the active tool pill, and the `selected_tool` request parameter are all untested. ## Proposed Cover these assertions: 1. Every tool configured in `canvas_dev_ai.settings` renders in the dropdown, with its label and description. 2. With no tools configured, t… ### [Config in language other than site default language causes an exception](https://git.drupalcode.org/project/canvas/-/work_items/3591850) State: opened | Updated: 2026-09-11 | Author: Ted Bowman (tedbow) | Assigned: unassigned | Labels: category::bug, state::needsWork, translation Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1400 [🤖 bug: #3591850 Use entity langcode, not site default, as default translation](https://git.drupalcode.org/project/canvas/-/merge_requests/1400) | State: opened | Author: Ted Bowman (tedbow), 809 diff lines Description: ### Problem When a Canvas config entity's `langcode` differs from the site default language, the Canvas editor throws exceptions. This happens when `PageRegion` (or `ContentTemplate`) entities are created while one language is the site default, and the site default is changed later. The entities keep their original `langcode`. Two exceptions: 1. `InvalidArgumentException: getTranslation() must not be called with the langcode of the default translation ("en")` — thrown when the preview validates a region with an auto-save draft. `CanvasConfigEntityTranslationsAreValidConstraintValidator` and… ### [Canvas AI: Proper chat history management in the backend](https://git.drupalcode.org/project/canvas/-/work_items/3592032) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: AKHIL BABU (AkhilBabu) | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai Description: ## Summary Before starting, read https://git.drupalcode.org/project/canvas/-/work_items/3591923#note_2377604. The dev chat does not keep the agent's history across turns. Each turn runs as a fresh agent: the stored agent state is deleted when the turn finishes, and the next turn rebuilds its history from the client transcript. ## Why Within one turn the backend owns a complete history — every tool call and every tool result — restored from the stored agent state on each hop. At the turn boundary that state is deleted, and the client transcript that replaces it carries message text only, so… ### [Canvas AI: Prevent the selected Tool from changing during a chat turn](https://git.drupalcode.org/project/canvas/-/work_items/3591952) State: closed | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint17, AI Initiative Sprint, ai, ai::inside Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1604 [#3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn](https://git.drupalcode.org/project/canvas/-/merge_requests/1604) | State: merged 2026-09-11 | Author: Ahmad Khalil (ahmad-khalil-imagex), 741 diff lines Description: ## Summary One chat turn in the dev AI chat runs as several requests to `/admin/api/canvas/ai-dev`, all sharing one `request_id`. Each of those requests carries the selected Tool as `selected_tool`. The Tools dropdown stays interactive throughout, so the user can change or clear the Tool while a turn is still running. The selection must be fixed for the whole turn, in the UI and on the server. Part of https://git.drupalcode.org/project/canvas/-/work_items/3591777. ## Current behavior A turn is not one request. Set the component agent as the main agent, open the chat with the browser networ… ### [Site template installation fails with Canvas 1.11](https://git.drupalcode.org/project/canvas/-/work_items/3592052) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: Description: ### Steps to reproduce Setup Drupal CMS with ddev ``` mkdir -p my-drupalcms-site && cd my-drupalcms-site ddev config --project-type=drupal11 --docroot=web ddev start ddev composer create-project drupal/cms ddev composer drupal:recipe-unpack ``` Verify that the canvas version is 1.11.0 ``` ddev composer show drupal/canvas ``` Now run the install command ``` ddev drush si recipes/drupal_cms_starter ``` The install fails with the following error message _`You have requested a synthetic service ("kernel"). The DIC does not know how to construct this service.`_ Now downgrade canvas ve… ### [Canvas AI: Unify error handling in `canvas_component_agent`'s tools](https://git.drupalcode.org/project/canvas/-/work_items/3592021) State: opened | Updated: 2026-09-11 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1626 [3592021: Add AiResponseValidator::formatErrors() to render accumulated tool...](https://git.drupalcode.org/project/canvas/-/merge_requests/1626) | State: opened | Author: Ahmad Khalil (ahmad-khalil-imagex), 978 diff lines Description: Apply the error-handling pattern from https://git.drupalcode.org/project/canvas/-/work_items/3591958 to the tools `canvas_component_agent` uses: the access check throws first, unchanged; every other check accumulates instead of stopping at the first error; and the accumulated errors are reported through `AiResponseValidator::formatErrors()`. See the "Proposed solution" section of https://git.drupalcode.org/project/canvas/-/work_items/3591958 for the full pattern, and `place_components`/`edit_components` (https://git.drupalcode.org/project/canvas/-/merge_requests/1584) for the reference implem… ### [Remove Canvas-only text format guidelines from non-Canvas forms](https://git.drupalcode.org/project/canvas/-/work_items/3592061) State: opened | Updated: 2026-09-10 | Author: John Locke (freelock) | Assigned: unassigned | Labels: category::bug, priority::minor, state::needsReview Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1637 [bug: #3592061 Remove Canvas-only text format guidelines from non-Canvas forms](https://git.drupalcode.org/project/canvas/-/merge_requests/1637) | State: opened | Author: John Locke (freelock), 133 diff lines Description: Canvas’s block and inline HTML guidelines appear on ordinary text-format fields even though those formats cannot be selected there. Only guidelines for the field’s available formats should appear. ## What is happening? On an anonymous comment form, three sets of instructions appear: - Filtered HTML - Drupal Canvas — Block HTML (locked) - Drupal Canvas — Inline HTML (locked) The form has no text-format selector, so Drupal’s guideline-switching JavaScript cannot hide the extra instructions. Canvas removes its exclusive formats from the selector options on non-Canvas forms, but leaves their … ### [<iframe sandbox> + `Window.postMessage()` to add defense in depth](https://git.drupalcode.org/project/canvas/-/work_items/3527515) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: JavaScript, Needs tests, Security improvements, category::task Description: >>> [!note] Migrated issue Reported by: [hooroomoo](https://www.drupal.org/user/3688872) >>> Overview https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/iframe#sandbox - Use the sandbox attribute on our current iframe usages - Convert any existing cross-frame communications to use postMessage() When the embedded document has the same origin as the embedding page, it is strongly discouraged to use both allow-scripts and allow-same-origin, as that lets the embedded document remove the sandbox attribute — making it no more secure than not using the sandbox attribute a… ### [Canvas AI: Update orchestrator for content-only / page-variant chrome](https://git.drupalcode.org/project/canvas/-/work_items/3592060) State: opened | Updated: 2026-09-10 | Author: Jordan Graham (COBadger) | Assigned: Jordan Graham (COBadger) | Labels: Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1636 [Draft: task(AI): #3592060 Canvas AI: Update orchestrator for content-only / page-variant chrome](https://git.drupalcode.org/project/canvas/-/merge_requests/1636) | State: opened | Author: Jordan Graham (COBadger), 150 diff lines Description: Problem: After !1624, page/template builders only fill content and refuse site chrome. The orchestrator still describes and routes as if they build headers/footers. **In scope (orchestrator install YAML + post-update to reimport it):** - Tool blurb: drop “including headers, footers…” - Routing: don’t send “add a header/footer” to page builder as if that places chrome - Capability list / examples that promise header/footer pages - Teach the right story: page variant supplies chrome; builders do body/content only (and what to say when the user asks for a header — e.g. explain variant chrome, d… ### [Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form](https://git.drupalcode.org/project/canvas/-/work_items/3592030) State: opened | Updated: 2026-09-10 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1624 [task(AI): #3592030 Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form](https://git.drupalcode.org/project/canvas/-/merge_requests/1624) | State: opened | Author: Ahmad Khalil (ahmad-khalil-imagex), 1352 diff lines Description: Canvas replaced theme-global regions with page variants. Two things follow from that: - The `content` region is unchanged — a page's own content is still always rendered inside it. - Global regions such as header, footer, and sidebar no longer exist. A page variant now lets a site builder decide what the rest of the page looks like — everything besides the main content area — as a single component tree, created and edited entirely from the Canvas editor UI. `canvas_ai` still has two things left over from the old multi-region model: - The `canvas_page_builder_agent` and `canvas_template_buil… ### [Normalize agent-written prop values against their schema and report values the component source cannot store](https://git.drupalcode.org/project/canvas/-/work_items/3592002) State: opened | Updated: 2026-09-10 | Author: Ahmad Khalil (ahmad-khalil-imagex) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative, AI Initiative Sprint, ai::inside Description: ### Problem/Motivation Split out of #3591921 (MR !1505) at the maintainer's request, see https://git.drupalcode.org/project/canvas/-/work_items/3591921#note_1979702 `AiResponseValidator::validateComponentStructure()` feeds the model's parsed YAML prop values straight into `clientModelToInput()` and the validation constraints. Two consequences show up in live builds with GPT-5.2: 1. **Equivalent textual shapes are rejected.** The model regularly writes `width: '"90%"'`, `level: "2"` or `section_header: "false"`. The YAML parser keeps the inner quotes / string type, every enum, integer and bo… ### [Canvas AI: Unify error handling in the dev page builder agent's tools](https://git.drupalcode.org/project/canvas/-/work_items/3591958) State: opened | Updated: 2026-09-10 | Author: AKHIL BABU (AkhilBabu) | Assigned: unassigned | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !1584 [Unify error handling in all tools used by the dev page builder agent.](https://git.drupalcode.org/project/canvas/-/merge_requests/1584) | State: opened | Author: AKHIL BABU (AkhilBabu), 1460 diff lines Description: **Scope narrowed from the original description.** This issue now covers only the four tools `canvas_dev_page_builder_agent` uses: `place_components`, `edit_components`, `set_page_value`, `get_component_details`. The remaining `canvas_ai` tools get their own follow-up issues, each pointing back to the proposed solution below instead of repeating it: https://git.drupalcode.org/project/canvas/-/work_items/3592021 (component agent), https://git.drupalcode.org/project/canvas/-/work_items/3592022 (live page builder agent's `SetAIGeneratedComponentStructure`), and https://git.drupalcode.org/project/c… ### [Add component label to settings form](https://git.drupalcode.org/project/canvas/-/work_items/3591934) State: opened | Updated: 2026-09-10 | Author: Andrew Morton (mortona2k) | Assigned: Lauri Timmanee (lauriii) | Labels: Needs design, state::needsReview Description: Sometimes the context of what you're editing can get lost. For example, if the selected component is scrolled out of view or the UI has glitched from a recent action, or delayed update. I've seen this happen when doing things quickly. A simple fix that would help is adding the component label to the top of the settings form. ### [The error message about "collapsed input syntax" could be more helpful](https://git.drupalcode.org/project/canvas/-/work_items/3558318) State: opened | Updated: 2026-09-10 | Author: drupalbot | Assigned: unassigned | Labels: category::bug, code, priority::normal, state::accepted Description: >>> [!note] Migrated issue Reported by: [phenaproxima](https://www.drupal.org/user/205645) >>> Overview In certain situations, you can receive this validation error: When using the default static prop source for a component input, you must use the collapsed input syntax. I think this message could be better: You can get this message for prop sources that aren't static, like host-entity-url. This is very confusing and can easily point you in the wrong direction -- I'd be inclined to start hunting for problems in my static prop sources, when in fact the problem is not coming from a static… --- MERGE REQUESTS (24h) --- ### [refactor(Conflict resolution): #3591785 Strengthen AutoSaveManager to generate stable hashes](https://git.drupalcode.org/project/canvas/-/merge_requests/1356) State: merged 2026-08-04 | Author: Feliksas Mazeikis (f.mazeikis) | Branch: 3591785-strengthen-autosave-normalization | 977 diff lines Description: ## Summary This branch consolidates the typed-data normalization logic that `AutoSaveManager` and `ComponentSourceBase` previously duplicated into a single method, `TypedDataHelper::castRawPhpTypes()`. ## What changed and why `AutoSaveManager::normalizeEntity()` (hash source) and `ComponentSourceBase::castRawTypedConfigToPhpTypes()` (component-shape normalization) each contained their own primitive-casting and traversal logic. These are merged into `TypedDataHelper::castRawPhpTypes()`, which already existed for the config typed-data path but was not wired into `toStorableArray()`. During c… ### [Draft: Resolve "Non HTML-validating raw template fails to add Canvas slot wrapper comments"](https://git.drupalcode.org/project/canvas/-/merge_requests/1638) State: opened | Author: dpi | Branch: 3569815-non-validating-raw | 320 diff lines Description: [Description of changes] Fully vibe coded, all vibes. Seems to work, pushed so we can share this for testing internally. Feel free to take the work and continue it though! ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3569815 ### [#3592031: Added padding to the container to fix visual gaps.](https://git.drupalcode.org/project/canvas/-/merge_requests/1600) State: opened | Author: utkarsh_33 (Utkarsh_33) | Branch: 3592031-brandkit-colors-ui-fixes | 194 diff lines Description: [Description of changes] ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3592031 ### [Draft: 🤖 feat: #3592009 Add locale-aware date/time formatting filters (Twig) and functions (JS)](https://git.drupalcode.org/project/canvas/-/merge_requests/1632) State: opened | Author: Narendra Singh Rathore (NarendraR) | Branch: 3592009-datetime-formatting-filters-and-functions | 1916 diff lines Description: [Description of changes] AI-Generated: Yes (Used Opencode (AI assistant, anthropic.claude-sonnet-4-6)). ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3592009 ### [Draft: Resolve "Support mapping multi-valued list fields to array-type component props"](https://git.drupalcode.org/project/canvas/-/merge_requests/1630) State: opened | Author: Chandan Singh (chandu7929) | Branch: 3592057-support-mapping-multi-valued-list-fields-to-array-type-component-props | 488 diff lines Description: This change adds support for mapping compatible multi-valued Drupal list fields to array-type Canvas component props. It updates three parts of Canvas: 1. **Validation** `MultiValuedFieldNotSupportedConstraintValidator` now allows supported list field types instead of rejecting all non-single-cardinality fields. This covers: - `list_string` - `list_integer` - `list_float` 2. **Field suggestion / picker flow** `ApiUiContentEntityReferenceControllers` now allows supported multi-valued scalar fields through the UI filtering logic, instead of excluding all multi-valued fields. 3. *… ### [Improve the Brand Kit Fonts UI](https://git.drupalcode.org/project/canvas/-/merge_requests/1520) State: opened | Author: Lauri Timmanee (lauriii) | Branch: 3591957-improve-the-brand-kit-fonts-ui | 3084 diff lines Description: Co-authored by Claude Code ## Testing instructions - [ ] First step for someone to test the changes in this MR Closes #3591957 ### [chore(deps): update dependency cspell to v10](https://git.drupalcode.org/project/canvas/-/merge_requests/1273) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/major-cspell-monorepo | 1082 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [cspell](https://cspell.org/) ([source](https://github.com/streetsidesoftware/cspell/tree/HEAD/packages/cspell)) | [`^9.2.0` → `^10.0.0`](https://renovatebot.com/diffs/npm/cspell/9.7.0/10.2.2) | ![age](https://developer.mend.io/api/mc/badges/age/npm/cspell/10.2.2?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/cspell/9.7.0/10.2.2?slim=true) | > ❗ **Important** >… ### [chore(deps): update dependency @eslint/compat to v2](https://git.drupalcode.org/project/canvas/-/merge_requests/1306) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/eslint-compat-2.x | 72 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [@eslint/compat](https://github.com/eslint/rewrite/tree/main/packages/compat#readme) ([source](https://github.com/eslint/rewrite/tree/HEAD/packages/compat)) | [`^1.3.1` → `^2.0.0`](https://renovatebot.com/diffs/npm/@eslint%2fcompat/1.4.1/2.1.1) | ![age](https://developer.mend.io/api/mc/badges/age/npm/@eslint%2fcompat/2.1.1?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confide… ### [chore(deps): update dependency @chromatic-com/storybook to v5](https://git.drupalcode.org/project/canvas/-/merge_requests/1305) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/chromatic-com-storybook-5.x | 13 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [@chromatic-com/storybook](https://github.com/chromaui/addon-visual-tests) | [`^3.2.2` → `^5.0.0`](https://renovatebot.com/diffs/npm/@chromatic-com%2fstorybook/3.2.7/5.3.1) | ![age](https://developer.mend.io/api/mc/badges/age/npm/@chromatic-com%2fstorybook/5.3.1?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/@chromatic-com%2fstorybook/3.2.7/5.3.1?slim=true) | >… ### [fix(deps): update npm minor and patch](https://git.drupalcode.org/project/canvas/-/merge_requests/1238) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/npm-minor-and-patch | 103 diff lines Description: This MR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [@axe-core/playwright](https://github.com/dequelabs/axe-core-npm) | [`4.11.1` → `4.13.0`](https://renovatebot.com/diffs/npm/@axe-core%2fplaywright/4.11.1/4.13.0) | ![age](https://developer.mend.io/api/mc/badges/age/npm/@axe-core%2fplaywright/4.13.0?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/@axe-core%2fplaywright/4.11.1/4.13.0?slim=true) | | [@babel/parser](… ### [chore(deps): update mysql docker tag to v8.3](https://git.drupalcode.org/project/canvas/-/merge_requests/1272) State: opened | Author: **** (project_19391_bot_cb443e60680723bba08a5351bf595251) | Branch: renovate/mysql-8.x | 13 diff lines Description: This MR contains the following updates: | Package | Type | Update | Change | Pending | |---|---|---|---|---| | [mysql](https://hub.docker.com/_/mysql) ([source](https://github.com/docker-library/mysql)) | service-image | minor | `8.0` → `8.3` | `8.4` | > ❗ **Important** > > Release Notes retrieval for this MR were skipped because no github.com credentials were available. > If you are self-hosted, please see [this instruction](https://github.com/renovatebot/renovate/blob/master/docs/usage/examples/self-hosting.md#githubcom-token-for-release-notes). --- ### Configuration 📅 **Schedule**: (UT… ### [Iframe sandbox: drop allow-same-origin, add postMessage defense in depth](https://git.drupalcode.org/project/canvas/-/merge_requests/1610) State: opened | Author: Matt Hand (matthand) | Branch: 3527515-iframe-sandbox-windowpostmessage-to-add-defense-in-depth | 2560 diff lines Description: ## Summary This merge request addresses a moderate security vulnerability by removing `allow-same-origin` from Canvas's sandboxed iframes—a combination with `allow-scripts` that represents "a well-known sandbox bypass." The changes affect two components: the code editor's live preview iframe and extension iframes. ## Code Editor Preview Changes The preview iframe has been restructured to eliminate regenerating `srcDoc` on each keystroke. Instead, it loads a fixed shell document once from a dedicated route with an opaque origin (`sandbox="allow-scripts"` only). Live updates now transmit via … ### [issue #3591965: add condition to just load lib on required theme](https://git.drupalcode.org/project/canvas/-/merge_requests/1529) State: opened | Author: julio retkwa (julioretkwa) | Branch: 3591965-canvas-library-is-loaded-on-admin-ui-even-when-not-requested | 14 diff lines Description: adds condition to just load `canvas/canvas.scoped.admin.css` if canvas_stark theme is the default ## Testing instructions * [ ] Install Drupal with the Gin admin theme active. * [ ] Install and enable the Canvas module * [ ] Create or edit a content type that includes a media library field (e.g. an image field using the Media Library widget). * [ ] Navigate to a node edit page for that content type * [ ] Click the "Add media" button to open the media library widget modal. * [ ] Observe that the Gin secondary toolbar (the bar containing "Search", "Shortcuts" ,"Drupal Canvas ") is displaced up… ### [feat(AI): #3592032 Canvas AI: Keep the agent history across chat turns in the backend](https://git.drupalcode.org/project/canvas/-/merge_requests/1635) State: opened | Author: Konstantine Kirkitadze (kotekirkitadze) | Branch: 3592032-issue-branch | 1821 diff lines Description: \## Problem Each dev-chat turn ran a fresh agent. The stored agent state was deleted when a turn finished, and the next turn was seeded from the client transcript, which carries text only. The agent lost every tool call and tool result of the conversation and repeated work, e.g. re-reading component schemas before every follow-up edit. \## Fix \- \`AiWizardDev\` generates one \`conversation_id\` per mount (the point where the chat history is cleared today) and sends it with every request, next to the per-turn \`request_id\`. \- \`CanvasAiTempStore\` gets a second key space, \`conversation_… ### [chore: #3592007 Allow developer and test tooling to locate the Drupal root when the module checkout lives outside the Drupal tree](https://git.drupalcode.org/project/canvas/-/merge_requests/1573) State: opened | Author: Bálint Kléri (balintbrews) | Branch: 3592007-allow-developer-and-test-tooling-to-locate-the-drupal-root-when-the-module-checkout-lives-outside-the-drupal-tree | 109 diff lines Description: ## Problem See #3592007. ## Proposal Honor `DRUPAL_TEST_DRUPAL_ROOT` as the absolute path to the Drupal web root, matching `@drupal/playwright` in all five problematic locations (see #3592007), keeping the existing searches as the fallback so behavior without the variable is unchanged: 1. `composer-root` resolves the project root from it (using `composer.json` beside or above the web root, the same idea `@drupal/playwright` applies). 1. `ui/cypress.config.js` derives the core directory as `$DRUPAL_TEST_DRUPAL_ROOT/core`; the existing `DRUPAL_ROOT_CORE` override keeps working and takes prec… ### [Sync brand kit colors through canvas.brand-kit.json in the Canvas CLI](https://git.drupalcode.org/project/canvas/-/merge_requests/1586) State: opened | Author: Lauri Timmanee (lauriii) | Branch: 3592019-sync-brand-kit-colors-through-canvasbrand-kitjson-in-the-canvas-cli | 5121 diff lines Description: Co-authored with Claude Code. Closes #3592019 ### [fix(Data model): #3558318 The error message about "collapsed input syntax" could be more helpful](https://git.drupalcode.org/project/canvas/-/merge_requests/1633) State: opened | Author: Georgi Ivanov (givanov) | Branch: 3558318-the-error-message-about-collapsed-input-syntax-could-be-more-helpful | 203 diff lines Description: ## Problem `ValidComponentTreeItemConstraintValidator` reported: > When using the default static prop source for a component input, you must use the collapsed input syntax. Two problems: 1. The violation fires for any prop source, not only static ones. A `host-entity-url` prop source stored as `{"sourceType": "host-entity-url"}` triggers it, because `absolute` defaults to `TRUE` and becomes explicit once collapsed. The message named the wrong cause. 2. "Collapsed input syntax" was defined in no docblock, comment or documentation. ## Solution The message now names the prop at fau… ### [Issue #3592055 by f0ns: Hide auto-created folders whose Components are all disabled](https://git.drupalcode.org/project/canvas/-/merge_requests/1628) State: opened | Author: Fons Vandamme (f0ns) | Branch: 3592055-library-shows-empty-auto-created-folders-for-categories-whose-components-are-all-disabled | 378 diff lines Description: ## What this does `Component::postSave()` created a Folder for every new Component regardless of status. Core-provided Block Components are generated disabled, so each of their categories got a Folder that shows up empty in the Library. Folder config entities are keyed by UUID, so a site template can't target them with config actions. A Folder is now only created for a Component that is **enabled**. That covers both a Component created enabled and one switched on later. It is deliberately *not* the `FALSE → TRUE` transition alone, since SDCs are created enabled and never undergo that transit… ### [fix(CLI Tool): #3592059 Improve Canvas CLI compatibility with Canvas <1.11](https://git.drupalcode.org/project/canvas/-/merge_requests/1631) State: opened | Author: wotnak | Branch: 3592059-improve-canvas-cli-compatibility-with-canvas-111 | 654 diff lines Description: - move the page template support check before authentication - continue syncing pages and content templates when page templates are not support just with page template fields skipped - improve page template unsupported message to suggest using CLI 0.23 on older sites ## Testing instructions - try to push/pull with `@drupal-canvas/cli@0.24.0` on a Canvas 1.10 site it will fail with missing oauth scope errors - rebuild the CLI with the MR changes - try to push/pull on the same site but with the rebuild CLI, it should now display helpful warning about page templates not being supported and pull/… --- COMMITS (24h) --- - [4c5b55b9](https://git.drupalcode.org/project/canvas/-/commit/4c5b55b97b062174f729f7bb3fd12ad11eebddc1) task(AI): #3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn — Ahmad Khalil (2026-09-11T06:23:48.000+00:00) ======================================================================== ## [24] summariseModule:ai_context:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Context Control Center (CCC) (machine name: ai_context) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. Do not name any individual people in this section. No one was active enough this period to list. Format your response as an HTML fragment. Start with <h3>Context Control Center (CCC)</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [CCC Pre-rc1 QA: Context items settings testing](https://git.drupalcode.org/project/ai_context/-/work_items/3586496) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::task Description: ## Problem/motivation Child of #3586460+s. Context item settings control publish-by-default, revisions, subcontext hierarchy, conditional AI decisions, and the Scheduler link when that module is installed. ## Solution Run the attached script `3586496-context-items-settings.md`. Cover publishing defaults, hierarchy off, conditional AI on/off and the parent-decision cap, deep links, and Scheduler. Comment Pass / Fail / Blocked / N/A for each case. File one issue per bug and relate it here and to #3586460+s. [00-shared-setup.md](https://git.drupalcode.org/-/project/191993/uploads/80e34ffdf… ### [Disabling "Create new revisions by default" still creates a new revision](https://git.drupalcode.org/project/ai_context/-/work_items/3586515) State: opened | Updated: 2026-09-11 | Author: tushar1 (Tushar1) | Assigned: unassigned | Labels: Description: ## Description Disabling **Create new revisions by default** does not prevent a new revision from being generated when an existing AI Context Item is edited and saved. ## Steps to reproduce 1. Go to `/admin/config/ai/context/settings/items`. 2. Enable **Create new revisions by default**. 3. Save the configuration. 4. Edit an existing AI Context Item and make a content change. 5. Save the item and verify that a new revision is created. 6. Return to Context Items Settings. 7. Disable **Create new revisions by default**. 8. Save the configuration. 9. Edit the same AI Context Item and … ### [Update context consumer listing configuration summary column](https://git.drupalcode.org/project/ai_context/-/work_items/3586514) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: Kristen Pol (kepol) | Labels: 2026Sprint18, consider for 1.0, state::toBeTriaged Description: to include what subscriptions are set ### [Change Add AI Context Item button to Add Context Item](https://git.drupalcode.org/project/ai_context/-/work_items/3586513) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: Kristen Pol (kepol) | Labels: 2026Sprint18, AI Initiative Sprint, ai::internal, category::task Description: I keep forgetting to do this and it's been on my mental list for many months ### [Accessibility Testing: CCC overview dashboard](https://git.drupalcode.org/project/ai_context/-/work_items/3586434) State: opened | Updated: 2026-09-10 | Author: Kristen Pol (kepol) | Assigned: Kati Vera (kati.vera) | Labels: 2026Sprint18, AI Initiative Sprint, accessibility, ai::inside Description: ## Problem/motivation Parent: #3586433+s _**THIS ISSUE IS RESERVED FOR MEARRA (KATI) - PING KRISTEN IN DRUPAL SLACK YOU WANT TO WORK ON THIS AND AREN'T KATI :smile:**_ _**THIS ISSUE WAS GENERATED WITH GROK AND NEEDS HUMAN REVIEW BEFORE PROCEEDING.**_ This is the first-batch area **#2**. This is an **area audit**, not a confirmed bug. Run the checks below on the Overview dashboard and file each confirmed finding as its own follow-up (one WCAG success criterion per issue). Do not change code from reading the template or from AI review alone. CCC has had keyboard/DOM work and some automat… ### [CCC Pre-rc1 QA: Test context item creation and workflow](https://git.drupalcode.org/project/ai_context/-/work_items/3586491) State: closed | Updated: 2026-09-10 | Author: Kristen Pol (kepol) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::task Description: ## Problem/motivation Child of #3586460+s. Context items are the core CCC content. Editors must be able to create, edit, view, duplicate, revise, moderate, and delete items without data loss or broken listing/view pages. ## Solution Run the attached script `3586491-context-item-creation.md`. Cover add/edit/view/delete/duplicate, Purpose validation, listing, revisions, Global vs scopes, subcontext rules, moderation, scheduling, import, and access. Comment Pass / Fail / Blocked / N/A for each case. File one issue per bug and relate it here and to #3586460+s. [00-shared-setup.md](https://g… ### [CCC Pre-rc1 QA: Context scope testing](https://git.drupalcode.org/project/ai_context/-/work_items/3586492) State: opened | Updated: 2026-09-10 | Author: Kristen Pol (kepol) | Assigned: Tazo Vekua (tazo_vekua) | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::task Description: ## Problem/motivation Child of #3586460+s. Scope settings and per-item scope widgets decide what is eligible for selection. Bugs here silently drop or leak context. ## Solution Run the attached script `3586492-context-scope.md`. Cover the scope overview, enable/disable without wiping stored values, Language unavailable on a single-language site, Site Sections, Entity Types, Roles, Specific Entities, Taxonomy, and consumer subscriptions. Comment Pass / Fail / Blocked / N/A for each case. File one issue per bug and relate it here and to #3586460+s. [00-shared-setup.md](https://git.drupalc… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [25] summariseModule:ai_ckeditor:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI CKEditor (machine name: ai_ckeditor) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Kristen Pol (kepol)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI CKEditor</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections](https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848) State: opened | Updated: 2026-09-11 | Author: Kristen Pol (kepol) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, category::bug, code Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !16 [Issue #3615848: Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections.](https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16) | State: opened | Author: Kristen Pol (kepol), 916 diff lines Description: ## Problem Follow-up to: ai_context#3586510+s `AiRequest::doRequest()` attaches `entity_context` request metadata so subscribers such as AI Context can scope pushed context to the entity being edited. It only works for saved entities. 1. **Unsaved entities have no context at all.** `ai_ckeditor_form_alter()` puts `entity_type` and `id` into drupalSettings. On an add form the id is an empty string, and `validateEntityContext()` returns NULL when the id is empty, so nothing is sent. AI Context's Entity Types scope then cannot match, so a context item scoped to "Content: Pirate post" is never… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [26] summariseModule:ai_dashboard:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI Dashboard (machine name: ai_dashboard) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Rob Loach (robloach) [1]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI Dashboard</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [How often is the ecosystem data updated?](https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221) State: opened | Updated: 2026-09-10 | Author: Michael Anello (ultimike) | Assigned: unassigned | Labels: Description: I'm curious how often the data on https://ai-dashboard-41c651.pages.drupalcode.org/ecosystem/ is updated. thanks, -mike --- MERGE REQUESTS (24h) --- ### [Issue #3585221: Rebuild the ecosystem](https://git.drupalcode.org/project/ai_dashboard/-/merge_requests/34) State: merged 2026-09-10 | Author: Rob Loach (robloach) | Branch: 3585221-rebuild_ecosystem | 4616 diff lines Description: Refreshes ecosystem.json and ecosystem.md (234 modules, 38 recipes as of 2026-09-10), and fixes the crawler to retry raw socket timeouts. --- COMMITS (24h) --- - [dfddf5e5](https://git.drupalcode.org/project/ai_dashboard/-/commit/dfddf5e5b35971cb0c30a08922b2971359c113b4) Issue #3585221: Rebuild the ecosystem — Rob Loach (2026-09-10T16:16:17.000+00:00) ======================================================================== ## [27] summariseModule:field_widget_actions:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: Field Widget Actions (machine name: field_widget_actions) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Artem Dmitriiev (a.dmitriiev), Joonas Meriläinen (mErilainen), Nik LePage (niklp) [1]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>Field Widget Actions</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Target element lookup fails for select widgets](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828) State: opened | Updated: 2026-09-11 | Author: Jürgen Haas (jurgenhaas) | Assigned: unassigned | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai::inside Description: FieldWidgetActionBase::getTargetElement() appends the constant FORM_ELEMENT_PROPERTY ('value') to the triggering button's parents when a delta exists, and falls back to the hardcoded path field/widget/0/value otherwise. That only matches widgets whose input is a value child, like plain text fields. Select, radios, checkboxes, tagify and select2 widgets fail: core marks them multiple_values: TRUE, so WidgetBase::form() never builds delta keys, and this module's fieldWidgetSingleElementFormAlter() wraps them in a container with the input under the widget key. The lookup then returns an empty ar… ### [Handle empty "Set widget field value" gracefully (with helpful message)](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578827) State: opened | Updated: 2026-09-11 | Author: Jürgen Haas (jurgenhaas) | Assigned: Shashank Rai (shqshqnk) | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai::inside Description: This got originally reported in project/eca_field_widget_actions#3588905+ by @ultimike We came to the conclusion that this should be handled by the FWA framework, but we can't move the other issue over here. Therefore, opening a new one. ### [Bug when trying to "Remove action"?](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578826) State: opened | Updated: 2026-09-11 | Author: Michael Anello (ultimike) | Assigned: unassigned | Labels: 2026Sprint17, 2026Sprint18, AI Initiative Sprint, ai::inside Description: I have a fresh Drupal 11.4.5 site with AI 1.4.7 and Field Widget Actions 1.4.1 After creating a simple "LLM: text" automator and adding it to a Text (plain, long) field via FWA, it works as expected - no problems there. But, if I return to the "Manage form display" page for the content type and try to remove the FWA from the field's configuration, it doesn't let me. Here's exactly what what I did and what happens: 1. `drush cr` (because, of course I did) 2. Navigate to the "Manage form display" page for the content type. 3. Click to configure the Text field (via its gear icon.) 4. Opened th… ### [Configured actions are invisible in the "Manage form display" summary row](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578830) State: closed | Updated: 2026-09-10 | Author: Nik LePage (niklp) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::feature Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !32 [Add action summary to manage fields ui](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/32) | State: merged 2026-09-10 | Author: Nik LePage (niklp), 45 diff lines Description: Created with the help of AI (CC) When a field widget has one or more actions configured (e.g. "Draft with AI"), that's only visible by opening the widget's settings (the gear icon) and expanding the "Field Widget Actions" details. The collapsed summary row on Manage form display — the same line that shows things like "Number of rows: 8" for a textarea — gives no indication that any actions are attached at all. This makes it easy to lose track of which fields have actions configured, especially on forms with many fields, since you have to open every field's settings individually to check. Dr… ### [TranslatableMarkup returned as string causes fatal error](https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578829) State: closed | Updated: 2026-09-10 | Author: Nik LePage (niklp) | Assigned: unassigned | Labels: 2026Sprint18, AI Initiative Sprint, ai::inside, category::bug Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !33 [#3578829: Cast return value to string in getDescription()](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/33) | State: merged 2026-09-10 | Author: Joonas Meriläinen (mErilainen), 13 diff lines Description: Created with the help of AI Current dev branch (src/FieldWidgetActionBase.php:106-108): ``` public function getDescription(): string { return $this->getPluginDefinition()['description'] ?? ''; } ``` The attribute declares $description as ?TranslatableMarkup (Attribute/FieldWidgetAction.php), so any non-null description returns a TranslatableMarkup object under the string return type — which causes a fatal. The `?? ''` only guards the null/missing case, not the type mismatch. Fix: one-line cast — `return (string) ($this->getPluginDefinition()['description'] ?? '');` --- MERGE REQUESTS (24h) --- ### [Add action summary to manage fields ui](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/35) State: merged 2026-09-10 | Author: Artem Dmitriiev (a.dmitriiev) | Branch: cherry-pick-ea54b3d2 | 45 diff lines Description: (cherry picked from commit ea54b3d201221ef49ee4ce0d9bc245c052c65320) 97097c30 Add action summary to manage fields ui 24ae1493 Fixed coding standards error f1ecf37d Fix coding standards Co-authored-by: Nik LePage ### [#3578829: Cast return value to string in getDescription()](https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/34) State: merged 2026-09-10 | Author: Artem Dmitriiev (a.dmitriiev) | Branch: cherry-pick-76fdceae | 13 diff lines Description: (cherry picked from commit 76fdceaed36a5c606bd64ff5185a7c5942945554) 76fdceae #3578829: Cast return value to string in getDescription() Co-authored-by: Joonas Meriläinen --- COMMITS (24h) --- - [ea54b3d2](https://git.drupalcode.org/project/field_widget_actions/-/commit/ea54b3d201221ef49ee4ce0d9bc245c052c65320) Add action summary to manage fields ui — Nik LePage (2026-09-10T13:57:15.000+00:00) ======================================================================== ## [28] summariseModule:ai_translate:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: AI translate (machine name: ai_translate) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Sven Decabooter (svendecabooter)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>AI translate</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [HTML markup is corrupted/not preserved when translating formatted text fields (text_with_summary, text_long)](https://git.drupalcode.org/project/ai_translate/-/work_items/3585525) State: opened | Updated: 2026-09-10 | Author: Carlos Busto Capeáns (cbccharlie) | Assigned: unassigned | Labels: automation::suggestionExists, category::bug, needs port to 2.x, priority::normal Description: ## Problem/Motivation When using AI Translate to translate content that contains formatted text fields (`text_with_summary`, `text_long`) with HTML markup (e.g. ``, ``, ``, ``, etc.), the translated output does not preserve the original HTML structure. The AI model either strips, alters, or partially reconstructs the HTML tags, resulting in broken or invalid markup stored in the translated field value. This affects any content type that uses fields with a text format such as "Full HTML" or "Basic HTML" via CKEditor. ## Steps to reproduce 1. Install and configure the `ai_translate` submodul… ### [Regression: Layout Builder content is no longer translated](https://git.drupalcode.org/project/ai_translate/-/work_items/3585553) State: opened | Updated: 2026-09-10 | Author: Sven Decabooter (svendecabooter) | Assigned: unassigned | Labels: category::bug, code, needs backport to 1.3.x, needs port to 2.x Description: ## Summary This occurs since the changes introduced by https://www.drupal.org/sa-contrib-2026-121 `TextExtractor::shouldExtract()` calls `access('view')` on every field, and core's `LayoutSectionItemList::defaultAccess()` returns `AccessResult::forbidden()` unconditionally, so `layout_builder__layout` fails that check for every account, including user 1 with `configure any layout`. The layout field is therefore dropped before `LbFieldExtractor` ever runs, and Layout Builder inline block content is never translated. The check was added in 1.3.2 and 1.4.1; the failure is silent, so translation… --- MERGE REQUESTS (24h) --- ### [Resolve "Regression: Layout Builder content is no longer translated"](https://git.drupalcode.org/project/ai_translate/-/merge_requests/49) State: opened | Author: Sven Decabooter (svendecabooter) | Branch: 3585553-issue-branch-1.4 | 708 diff lines Description: ## Description Closes #3585553. `TextExtractor::shouldExtract()` checks `access('view')` on every field, and core's `LayoutSectionItemList::defaultAccess()` returns `AccessResult::forbidden()` for every account, including user 1. So `layout_builder__layout` never passes that check, the layout field is dropped before `LbFieldExtractor` is reached, and Layout Builder inline blocks are never translated. The translation still saves and reports success, so the failure is silent. Affects 1.3.2 and 1.4.1, with both `layout_builder_at` and `layout_builder_st`. The fix adds `ContainerFieldTextExtrac… --- COMMITS (24h) --- (none) ======================================================================== ## [29] summariseModule:elevenlabs:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: ElevenLabs (machine name: elevenlabs) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Anne Bonham (annebonham)". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>ElevenLabs</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Text-to-speech audio is truncated to the first ~5,000 characters because MP3 chunks are joined by raw byte concatenation](https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581) State: opened | Updated: 2026-09-11 | Author: Anne Bonham (annebonham) | Assigned: unassigned | Labels: Related MRs (already exist for this issue — check state before suggesting new contributions): - MR !5 [Issue #3597581 by banoodle: Fix audio truncated to the first chunk when text...](https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5) | State: opened | Author: Anne Bonham (annebonham), 282 diff lines Description: **Problem/Motivation** On a site where I have this enabled, the client noticed that the ElevenLabs-generated audio files were playing only the first 5,000 characters of text, even though the Elevenlabs limit is 10,000 characters.\ \ `ElevenlabsProvider::textToSpeech()` splits text longer than 5,000 characters into\ chunks, requests one MP3 per chunk, and then joins the results. The join is a raw\ byte append:\ \ `audio_binary_data = _NULL_;\ foreach (``audio_files as $audio_file) {$audio_binary_data .= $audio_file->getBinary();}`\ \ Every response from the ElevenLabs API is a complete MP3 fil… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- (none) ======================================================================== ## [30] summariseModule:flowdrop:executive ======================================================================== You are a technical writer producing a newsletter section about recent Drupal module activity. Module: FlowDrop (machine name: flowdrop) Reporting period: last 24h (2026-09-10T12:08:50+00:00 to 2026-09-11T12:08:50+00:00) You are writing for a non-technical executive audience (CEO/leadership level). Focus on: business impact, strategic progress, risks, and what is being delivered. Avoid technical jargon. Do not mention branch names, function names, or API details. Explain what each piece of work means for users or the project's goals. Keep risk language proportionate to the stage of the software. For alpha or beta modules, unassigned bugs or open issues are normal and expected — do not describe them as 'reputational risks' or use alarmist framing. Reserve strong risk language for issues that are genuinely critical or blocking a stable release. Focus your report on activity that occurred within the reporting period (comments, MRs merged, commits pushed, status changes). Use the issue description to understand what each issue is about and provide context, but do not report on the description itself as news — it is background information only. Do not list every issue/MR individually — synthesise into prose. Keep it under 200 words. Do not use emoticons or mdashes. Do not wrap usernames or contributor names in <strong> tags — mention them as plain text. When mentioning a specific issue or MR, always hyperlink it using the URL provided in the data (e.g. <a href="URL">Issue Title</a> or the Markdown equivalent). Do not reference issues or MRs by number alone — always use their title as the link text. Do NOT attribute any action to a named individual. Do not say who merged, fixed, reviewed, authored, or opened anything. The "Author" and "Assigned" labels in the data are context only — never surface them as "X did Y". Describe the work itself in the passive voice (e.g. "the provider refactor was merged"), not the person who did it. The author of a merge request is often not the person who completed or merged the work, so naming them is misleading. End your section with a single line listing the people who were active this period, exactly in this form (plain text, no <strong>): "Contributors: Shibin Das [8]". Use only the names from that list, verbatim, keeping any trailing "[N]" commit count exactly as shown. Do not introduce any other names. Format your response as an HTML fragment. Start with <h3>FlowDrop</h3> then use <h4>, <p>, <ul>/<li>, and <strong> as needed. Output only the HTML fragment with no surrounding <html>, <body>, or <article> tags. After the project summary prose, add a single subsection titled "<h4>How can I help on this project?</h4>" aimed at a non-technical executive. Suggest 2-3 concrete, high-level ways a leader could support or unblock progress (e.g. resourcing, stakeholder alignment, decision-making, funding, advocacy). Keep it under 60 words. Do not add any other 'How can I help' text anywhere else in the section. CRITICAL: Before writing each suggestion, verify it against the issue data. Do not suggest actions that are already in progress or covered by an existing MR — for example, do not suggest that work needs to be started if an issue already has a Related MR. --- ISSUES UPDATED IN LAST 24h --- Each issue includes its description (for context) and the most recent comments from the last 24h (or last 5 if fewer than 5 occurred in that window). ### [Bare editor/playground page leaks <head-placeholder>, pushing every stylesheet into <body> so UI themes (Drafter, Minimal) and skins never apply](https://git.drupalcode.org/project/flowdrop/-/work_items/3592452) State: opened | Updated: 2026-09-10 | Author: Shibin Das (D34dMan) | Assigned: unassigned | Labels: bug Description: ## Problem `BarePageRenderer` (flowdrop_ui_components) emits the literal tag `` in its HTML template. Drupal core's `HtmlResponseAttachmentsProcessor` only replaces the head placeholder when the response attaches a non-empty `html_head`. The bare page attaches none, so the raw tag is shipped to the browser. An unknown element is not allowed inside ``, so the HTML parser closes `` at that point and everything after it — the ``, the inline reset `` and **every ``** — is parsed into ``, wrapped in a stray `` element. ## Effect The `@flowdrop/flowdrop` library applies a UI theme (Drafter, Mini… --- MERGE REQUESTS (24h) --- (none) --- COMMITS (24h) --- - [2c459ecc](https://git.drupalcode.org/project/flowdrop/-/commit/2c459eccb3ece34235e0d5e3a4e9bc71556d9282) ci: exclude the vendored WebMCP bridge from eslint, prettier and cspell; add four dictionary words — Shibin Das (2026-09-10T23:58:54.000+02:00) - [49666c5f](https://git.drupalcode.org/project/flowdrop/-/commit/49666c5f14d7138d6e70d880672ab29cc716792f) chore: re-stamp 2.6.0-alpha1 with the chat tool loop — Shibin Das (2026-09-10T23:23:44.000+02:00) - [fca1f985](https://git.drupalcode.org/project/flowdrop/-/commit/fca1f98516e5e8898d12d2581bbe7c16b9aed7a6) docs: CHANGELOG — Unreleased entries for the chat tool loop, the editor surface switches, the bare page head fix and the 2.8.0 bundle — Shibin Das (2026-09-10T23:13:19.000+02:00) - [98717d50](https://git.drupalcode.org/project/flowdrop/-/commit/98717d5026854f8051bf5790f88d3d7ef66c0e2c) fix(flowdrop_chat): #3592451 a tool-heavy turn no longer wedges on the memory cap — Shibin Das (2026-09-10T23:13:19.000+02:00) - [0f351c9c](https://git.drupalcode.org/project/flowdrop/-/commit/0f351c9cedd4a94691ddf6eae12cead1b03d6a0c) test(flowdrop_ui_components): SurfaceSettingsTest no longer installs flowdrop_chat — Shibin Das (2026-09-10T23:13:19.000+02:00) - [e1eb94c2](https://git.drupalcode.org/project/flowdrop/-/commit/e1eb94c2800ed38ed9f2a5351254ed45356ab1ec) chore: bump flowdrop js lib to 2.8.0 — Shibin Das (2026-09-10T23:05:11.000+02:00) - [5a4a2b69](https://git.drupalcode.org/project/flowdrop/-/commit/5a4a2b69b714c56ff31b7b0df399eacbfc311d6d) feat(flowdrop_ui_components): site switches for the Command Console and the AI Assistant; no dead Sync to Cloud — Shibin Das (2026-09-10T16:36:01.000+02:00) - [e8e5dc4c](https://git.drupalcode.org/project/flowdrop/-/commit/e8e5dc4c0265a82f2648176959196414e05672ca) fix(flowdrop_ui_components): #3592452 bare page leaks head placeholder, stylesheets land in body — Shibin Das (2026-09-10T15:45:56.000+02:00) ======================================================================== ## [31] generateTldr:executive ======================================================================== You are an editor distilling a Drupal AI project newsletter into its most important highlights. You are writing for a non-technical executive audience. Focus on business impact, strategic progress, and delivery milestones. Avoid all technical jargon. Read all the module summaries below. Separate the highlights into two categories: - SHIPPED: things that were merged, fixed, released, or completed during this period. - ONGOING: things that are actively in progress, under review, or blocked. Be specific — name the module, what happened, and why it matters. Do not use emoticons or mdashes. Do not include any text outside the two sections. When you mention a specific issue or merge request, add an inline citation using [N] where N is the reference number from the list below. You may cite multiple references per item, e.g. [1][3]. Place citations immediately after the relevant phrase, before any punctuation. Format as two HTML sections. Use exactly this structure (all <li> elements must be inside the <ol>, never outside it): <h4>Shipped</h4> <ol> <li><strong>Title here</strong> — One sentence explanation [1].</li> <li><strong>Another title</strong> — One sentence explanation.</li> </ol> <h4>Ongoing</h4> <ol> <li><strong>Title here</strong> — One sentence explanation [2].</li> </ol> Up to 5 items per section. Do not output any text, tags, or characters outside these two sections. Output only the HTML fragment, no surrounding tags. --- REFERENCE LIST --- [1] LLM: Metatag generates and stores tags that were never configured, and keeps only one of them (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586723 [2] Allow Field Widget Actions to target a specific automator 2.x (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586657 [3] Review AI issue process and tagging based on recent usage and update docs (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3550025 [4] Create AI assistant API session only after the first message (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3554797 [5] Add interactive refinement capability to Field Widget Actions 2.x (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586655 [6] Automator FWA buttons give no user feedback when values don't visibly change (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3585987 [7] Restrict access to the AI API Explorer permission (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586733 [8] SetupVdbServer and SetupVdbIndex config actions aren't idempotent, recipe re-apply fails (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586729 [9] Adding Disable Option for AI Automators (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3508800 [10] Bug — the "Content Entity Seeder" tool generates a JSON Schema with one extra array level and breaks when creating/editing content (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586732 [11] textToImage() base64-decodes the image payload twice (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586730 [12] Regression in ai_translate: Layout Builder content is no longer translated (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586731 [13] Streaming chat forces plain-text hostname filtering, corrupting URLs inside HTML attributes (e.g., href="") (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586624 [14] Issue #3586657: Allow Field Widget Actions to target a specific automator 2.x (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1999 [15] Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x" (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2002 [16] feat: #3585987 by petar.basic: Add status message after AI Field Widget Action automator runs. (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1540 [17] Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x" (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1966 [18] Draft: fix: Fix AI Settings form to display AI Platforms and Inference Providers (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1901 [19] fix: #3586729 Add existence checks to `SetupVdbServer` and `SetupVdbIndex` config actions (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2000 [20] bug: #3586730 textToImage() base64-decodes the image payload twice (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2001 [21] Support media entities for vision models (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289 [22] Stable 1.0.0 release for use in Drupal CMS recipes (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025 [23] Replace token references by roken replacements (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027 [24] fix: #3585027 Use token replacement for the token input and mask fields (ai_integration_eca, MR) — https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26 [25] Add missing OpenAI TTS voices to api_defaults.yml (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3522771 [26] Add test coverage for Whisper (speechToText) and TTS (textToSpeech) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894 [27] Support batched embeddings (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568659 [28] Moderation endpoint has hardcoded model (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418 [29] "Selected API Key not working" after update to AI 1.3.0 (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3579428 [30] Add reasoning output (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3574226 [31] Change the way that system prompt is being added in 1.2.x (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3573133 [32] Add AI usage checkboxes to ai_provider_openai issue template (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568620 [33] Add AI 2.x / Search AI compatible OpenAI embedding provider (2.x branch) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3564086 [34] Support File Handling (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3548526 [35] Remove OpenAI SDK dependency and extend it from Drupal AI core module (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531260 [36] Handle pdf files in chat (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3529603 [37] Version of modules can go out of alignment (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519419 [38] Abstract token usage support (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519302 [39] Usage stats in streaming chat need conversion to an array. (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486224 [40] Log flagged prompts (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3507407 [41] Moderation log overview (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3516044 [42] Streaming chat responses gives Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools() (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3515885 [43] Allow skipping of moderations for some embeddings (not all) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3510601 [44] Bug: Key validation and Rate Limit tester are bypassing host variable; causing Auth errors or Network errors, etc. (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3496412 [45] Selecting "o1" model on automator produces error: "Unsupported parameter: 'max_tokens' is not supported" (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3508042 [46] Error with OpenAI o1 models (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489201 [47] Moderation reads confidence scores instead of the actual violation flags, so almost every category shows as flagged (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161 [48] Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat() (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601 [49] Add support for fast mode (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211 [50] Unable to do GPT 5.4 calls (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729 [51] How to enable type: web_search in Chat? (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3613029 [52] Automated Drupal 12 compatibility fixes for ai_provider_openai 1.3.x-dev (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3606117 [53] Create a skill to setup the provider (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3605812 [54] Automated Drupal 12 compatibility fixes for ai_provider_openai 1.2.x-dev (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3594765 [55] Opt-in to cache embedding lookups (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3552523 [56] Turn off moderation if omni-moderation-latest is not available (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3534855 [57] Add gpt-image-1 image to image capabilitites (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531289 [58] Warning appears while using automators for image type field "Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() " (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3504849 [59] Add event triggering for streaming (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489846 [60] Improve documentation on setting key (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486487 [61] Enable Drupal AI Module Functionality for Environments Behind a Proxy (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3481664 [62] Reasoning models return an empty message, and model-specific parameters are rejected by Converse (ai_provider_aws_bedrock, Issue) — https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/work_items/3594715 [63] Issue #3594715: Read the first content block that carries text, and route model-specific parameters to additionalModelRequestFields (ai_provider_aws_bedrock, MR) — https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19 [64] Structured output: an invalid JSON schema is silently accepted instead of throwing an exception (ai_provider_azure, Issue) — https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013 [65] Add batch embeddings (ai_vdb_provider_postgres, Issue) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3610509 [66] indexItems() throws TypeError on array_diff() because $skipDeleteItemIds is never declared (ai_vdb_provider_postgres, Issue) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3614504 [67] Issue #3614504: Fix TypeError in indexItems() from undeclared $skipDeleteItemIds property (ai_vdb_provider_postgres, MR) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33 [68] feat: #3610509 Add batch embeddings (ai_vdb_provider_postgres, MR) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30 [69] Dependency Dashboard (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591657 [70] [>=11.4.7] Broaden our check to determine if a recipe is being applied, use 11.4.7's new `RecipeRunner::isApplying()` (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591985 [71] RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591996 [72] [PP-1] Add logic + docs to not ship Component config entities with Drupal extensions (modules, themes, recipes), and how to do it anyway when there's a good enough reason (pre-populated component trees in Canvas Pages or ContentTemplates) (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3565712 [73] Allow Canvas Patterns+PageRegions+ContentTemplates provided by Recipes to NOT specify component versions (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3571366 [74] Canvas AI: Add Playwright coverage for the chat Tools dropdown (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591951 [75] Config in language other than site default language causes an exception (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591850 [76] Canvas AI: Proper chat history management in the backend (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592032 [77] Canvas AI: Prevent the selected Tool from changing during a chat turn (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591952 [78] Site template installation fails with Canvas 1.11 (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592052 [79] Canvas AI: Unify error handling in `canvas_component_agent`'s tools (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592021 [80] Remove Canvas-only text format guidelines from non-Canvas forms (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592061 [81] <iframe sandbox> + `Window.postMessage()` to add defense in depth (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3527515 [82] Canvas AI: Update orchestrator for content-only / page-variant chrome (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592060 [83] Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592030 [84] Normalize agent-written prop values against their schema and report values the component source cannot store (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592002 [85] Canvas AI: Unify error handling in the dev page builder agent's tools (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591958 [86] Add component label to settings form (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591934 [87] The error message about "collapsed input syntax" could be more helpful (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3558318 [88] Check for the recipe runner (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1554 [89] refactor(Conflict resolution): #3591785 Strengthen AutoSaveManager to generate stable hashes (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1356 [90] Add docs/recipes.md: how Canvas Component versions interact with recipes... (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1634 [91] Draft: Resolve "Non HTML-validating raw template fails to add Canvas slot wrapper comments" (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1638 [92] #3592031: Added padding to the container to fix visual gaps. (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1600 [93] Draft: 🤖 feat: #3592009 Add locale-aware date/time formatting filters (Twig) and functions (JS) (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1632 [94] 🤖 bug: #3591850 Use entity langcode, not site default, as default translation (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1400 [95] Draft: Resolve "Support mapping multi-valued list fields to array-type component props" (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1630 [96] Improve the Brand Kit Fonts UI (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1520 [97] #3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1604 [98] chore(deps): update dependency cspell to v10 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1273 [99] chore(deps): update dependency @eslint/compat to v2 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1306 [100] chore(deps): update dependency @chromatic-com/storybook to v5 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1305 [101] fix(deps): update npm minor and patch (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1238 [102] chore(deps): update mysql docker tag to v8.3 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1272 [103] bug: #3592061 Remove Canvas-only text format guidelines from non-Canvas forms (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1637 [104] Iframe sandbox: drop allow-same-origin, add postMessage defense in depth (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1610 [105] Draft: task(AI): #3592060 Canvas AI: Update orchestrator for content-only / page-variant chrome (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1636 [106] task(AI): #3592030 Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1624 [107] 3592021: Add AiResponseValidator::formatErrors() to render accumulated tool... (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1626 [108] issue #3591965: add condition to just load lib on required theme (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1529 [109] test(AI): #3591951 Canvas AI: Add Playwright coverage for the chat Tools dropdown (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1607 [110] feat(AI): #3592032 Canvas AI: Keep the agent history across chat turns in the backend (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1635 [111] Unify error handling in all tools used by the dev page builder agent. (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1584 [112] chore: #3592007 Allow developer and test tooling to locate the Drupal root when the module checkout lives outside the Drupal tree (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1573 [113] Sync brand kit colors through canvas.brand-kit.json in the Canvas CLI (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1586 [114] fix(Data model): #3558318 The error message about "collapsed input syntax" could be more helpful (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1633 [115] Issue #3592055 by f0ns: Hide auto-created folders whose Components are all disabled (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1628 [116] fix(CLI Tool): #3592059 Improve Canvas CLI compatibility with Canvas <1.11 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1631 [117] CCC Pre-rc1 QA: Context items settings testing (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586496 [118] Disabling "Create new revisions by default" still creates a new revision (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586515 [119] Update context consumer listing configuration summary column (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586514 [120] Change Add AI Context Item button to Add Context Item (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586513 [121] Accessibility Testing: CCC overview dashboard (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586434 [122] CCC Pre-rc1 QA: Test context item creation and workflow (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586491 [123] CCC Pre-rc1 QA: Context scope testing (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586492 [124] Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections (ai_ckeditor, Issue) — https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848 [125] Issue #3615848: Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections. (ai_ckeditor, MR) — https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16 [126] How often is the ecosystem data updated? (ai_dashboard, Issue) — https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221 [127] Issue #3585221: Rebuild the ecosystem (ai_dashboard, MR) — https://git.drupalcode.org/project/ai_dashboard/-/merge_requests/34 [128] Target element lookup fails for select widgets (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828 [129] Handle empty "Set widget field value" gracefully (with helpful message) (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578827 [130] Bug when trying to "Remove action"? (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578826 [131] Configured actions are invisible in the "Manage form display" summary row (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578830 [132] TranslatableMarkup returned as string causes fatal error (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578829 [133] Add action summary to manage fields ui (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/35 [134] #3578829: Cast return value to string in getDescription() (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/34 [135] Add action summary to manage fields ui (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/32 [136] #3578829: Cast return value to string in getDescription() (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/33 [137] HTML markup is corrupted/not preserved when translating formatted text fields (text_with_summary, text_long) (ai_translate, Issue) — https://git.drupalcode.org/project/ai_translate/-/work_items/3585525 [138] Regression: Layout Builder content is no longer translated (ai_translate, Issue) — https://git.drupalcode.org/project/ai_translate/-/work_items/3585553 [139] Resolve "Regression: Layout Builder content is no longer translated" (ai_translate, MR) — https://git.drupalcode.org/project/ai_translate/-/merge_requests/49 [140] Text-to-speech audio is truncated to the first ~5,000 characters because MP3 chunks are joined by raw byte concatenation (elevenlabs, Issue) — https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581 [141] Issue #3597581 by banoodle: Fix audio truncated to the first chunk when text... (elevenlabs, MR) — https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5 [142] Bare editor/playground page leaks <head-placeholder>, pushing every stylesheet into <body> so UI themes (Drafter, Minimal) and skins never apply (flowdrop, Issue) — https://git.drupalcode.org/project/flowdrop/-/work_items/3592452 --- MODULE SUMMARIES --- <h3>AI (Artificial Intelligence)</h3> <p>Development activity this period focused on stabilizing Field Widget Actions (FWA) capabilities and addressing defects in several automators and providers. Interactive refinement features for Field Widget Actions were successfully integrated into the 2.x version, giving editors the ability to iteratively improve AI-generated content within forms. A parallel effort to allow FWA buttons to target specific automators rather than running all configured automators simultaneously is now under review.</p> <p>Several bugs were identified and are being corrected. A critical <a href="https://git.drupalcode.org/project/ai/-/work_items/3586730">image generation defect</a> where images were corrupted due to double-decoding was addressed, preventing silent data corruption in generated files. A <a href="https://git.drupalcode.org/project/ai/-/work_items/3586729">configuration idempotency issue</a> affecting recipe-based deployments was also reported and now has a fix under review. Additional bugs surfaced in the Metatag automator (storing all tags rather than only configured ones), streaming chat (corrupting URLs in HTML attributes), and content translation for Layout Builder pages.</p> <p>One <a href="https://git.drupalcode.org/project/ai/-/work_items/3586733">security permission</a> issue has been flagged as a release blocker: the AI API Explorer tool's permission is not marked security-sensitive, despite allowing direct invocation of AI providers and function calls outside normal access controls.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize review and testing of the release blocker permission fix to ensure secure deployment readiness.</li> <li>Allocate resources to address the critical Layout Builder translation regression impacting multilingual content workflows.</li> </ul> <p>Contributors: Abhisek Mazumdar (<a href="https://www.drupal.org/u/abhisekmazumdar">abhisekmazumdar</a>), Artem Dmitriiev (<a href="https://www.drupal.org/u/a.dmitriiev">a.dmitriiev</a>), Dezső Biczó (<a href="https://www.drupal.org/u/mxr576">mxr576</a>), Juan Correa (<a href="https://www.drupal.org/u/jucs7">jucs7</a>), Petar Bašić (<a href="https://www.drupal.org/u/petar_basic">petar_basic</a>), Sven Decabooter (<a href="https://www.drupal.org/u/svendecabooter">svendecabooter</a>)</p> --- <h3>AI Integration - ECA</h3> <p>The AI Integration - ECA module progressed significantly toward production readiness during this period. Three issues were closed, including the completion of <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289">Support media entities for vision models</a>, which enables the module to process image and media files for AI-powered categorization and analysis. This expands the module's capability beyond text-only prompts, allowing organizations to automate visual content workflows.</p> <p>A critical stability milestone was achieved with the closure of <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025">Stable 1.0.0 release for use in Drupal CMS recipes</a>. The module is now production-ready and compatible with Drupal CMS recipe requirements, removing a packaging barrier that prevented adoption by products requiring stable releases. With 934 sites currently using the module, this opens the door to broader distribution.</p> <p>Internal consistency was improved through <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027">Replace token references by token replacements</a>, which standardizes how dynamic content is handled across the module's features, building on patterns established in the media support work.</p> <h4>How can I help on this project?</h4> <ul> <li>Promote the 1.0.0 stable release to downstream teams and partners building Drupal CMS recipes or packaged products.</li> <li>Consider case studies or documentation showcasing vision model use cases to drive adoption of the new media capabilities.</li> </ul> <p>Contributors: Jürgen Haas (<a href="https://www.drupal.org/u/jurgenhaas">jurgenhaas</a>) [1]</p> --- <h3>OpenAI Provider</h3> <p>The OpenAI Provider module saw primarily administrative activity over the last 24 hours, with 23 issues closed. These closures reflect past work being finalized rather than new development landing in the reporting period. The closed issues span multiple areas: support for additional text-to-speech voices, batched embeddings for faster vector database indexing, reasoning output from AI models, file handling capabilities, moderation logging interfaces, and compatibility fixes for various OpenAI model parameters. This cleanup work brings the project's issue queue into better alignment with its current functionality.</p> <p>Several active issues remain open and require attention. A <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161">moderation accuracy bug</a> is causing almost every content category to appear flagged because the system reads confidence scores instead of actual violation flags. An <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601">integration error</a> prevents all chat operations when using version 1.3.0-beta2 with the stable AI module. Two feature requests relate to supporting OpenAI's newer capabilities: <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211">fast mode</a> for improved response times and <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729">GPT 5.4 compatibility</a>.</p> <p>The project has 42 open issues in total, including work underway to improve test coverage for speech-to-text and text-to-speech operations. This is expected and appropriate for a module in active development supporting a rapidly evolving external API.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize the moderation accuracy bug and chat integration error, as these block core functionality for users on current stable versions.</li> <li>Clarify the module's roadmap for supporting new OpenAI features like fast mode and GPT 5.4, particularly whether these require upstream AI framework changes.</li> </ul> --- <h3>AWS Bedrock Provider</h3> <p>A bug affecting reasoning models was identified and a fix has been prepared. When using reasoning models, the system was returning empty responses instead of the actual generated text. This occurred because reasoning models structure their output differently, placing reasoning content in a separate block before the main response. The current code was looking in the wrong location for the text.</p> <p>The same issue affects the handling of model-specific configuration parameters, which were being rejected instead of properly routed to the underlying service.</p> <p>A <a href="https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19">fix for both problems</a> is now under review. Once merged, reasoning models will work correctly and model-specific parameters will be properly supported. This expands the module's compatibility with advanced AI capabilities and gives users more control over model behavior.</p> <h4>How can I help on this project?</h4> <ul> <li>Ensure testing resources are available to validate the reasoning model fix before release</li> <li>Consider prioritizing review capacity to move the pending fix through the approval process</li> </ul> Contributors: Besart Bytyqi (<a href="https://www.drupal.org/u/besart">besart</a>) --- <h3>Microsoft Azure AI</h3> <p>The Microsoft Azure AI provider module saw issue reporting activity but no code changes during this period. A <a href="https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013">validation gap in structured output handling</a> was documented, where invalid JSON schemas are silently accepted rather than flagged. This means that configuration errors—such as marking fields as required without defining them—are not caught early, potentially leading to unexpected behavior downstream. The issue is open and unassigned, with no fix currently in development.</p> <p>This represents normal maintenance activity for a module in active development. The issue has been logged and is awaiting prioritization and assignment.</p> <h4>How can I help on this project?</h4> <ul> <li>Allocate developer time to triage and assign the open validation issue, ensuring it is prioritized appropriately within the roadmap.</li> <li>Consider resourcing a quality assurance review of schema validation across the module to identify similar gaps proactively.</li> </ul> --- <h3>Postgres VDB Provider</h3> <p>The Postgres VDB Provider is being adapted to support batch embeddings, a new capability that enables indexing systems to process and store multiple content chunks in a single request instead of one at a time. This change is expected to meaningfully reduce indexing time for sites using this provider. The work aligns the Postgres provider with improvements already made to other vector database providers in the AI module ecosystem. A merge request implementing this capability is now open for review.</p> <p>Separately, a bug was identified and addressed that prevented any content indexing from working at all. When attempting to index content through Search API, the system would fail with an error due to an uninitialized variable. A fix has been prepared and is awaiting merge.</p> <p>These updates move the module toward more efficient operation and restore core indexing functionality that was broken.</p> <h4>How can I help on this project?</h4> <ul> <li>Ensure reviewer capacity is available to evaluate and merge the two open merge requests so the performance improvements and bug fix reach users promptly.</li> <li>Consider prioritizing resources for testing the batch embeddings feature in realistic production scenarios to validate the performance gains.</li> </ul> Contributors: Sergiu Nagailic (<a href="https://www.drupal.org/u/Nikro">Nikro</a>), Zeeshaan khann (<a href="https://www.drupal.org/u/zeeshan_khan">zeeshan_khan</a>) --- <h3>Drupal Canvas</h3> <h4>What shipped</h4> <p>A significant fix to the AI chat experience was completed: the <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591952">selected Tool can no longer change mid-turn</a>. Previously, because each chat turn involved multiple backend requests, users could inadvertently switch or clear the AI tool while a conversation was still running, causing inconsistent behavior. The system now locks the tool selection for the duration of a turn, both in the interface and on the server.</p> <h4>In progress</h4> <p>Work continues on several fronts. An <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591996">installation failure affecting site templates</a> was identified and is under review—the issue causes a container error when applying recipes with Canvas 1.11. Separately, efforts are underway to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592030">remove outdated AI instructions tied to the old region-based layout model</a>, now replaced by page variants. Progress is also being made on <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591958">unifying error handling across AI agent tools</a> and improving <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592002">validation of AI-generated content</a> to reduce unnecessary rejections. On the developer experience side, work is moving forward to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592007">support module development outside the Drupal tree</a> and to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592019">sync brand kit colors through the Canvas CLI</a>.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize diagnosis and resolution of the site template installation failure, which may block adoption of Canvas 1.11 in production environments.</li> <li>Ensure developer tooling improvements receive adequate review and testing resources to accelerate contributor onboarding.</li> </ul> <p>Contributors: **** (<a href="https://www.drupal.org/u/project_19391_bot_cb443e60680723bba08a5351bf595251">project_19391_bot_cb443e60680723bba08a5351bf595251</a>), Adam G-H (<a href="https://www.drupal.org/u/phenaproxima">phenaproxima</a>), Ahmad Khalil (<a href="https://www.drupal.org/u/ahmad-khalil-imagex">ahmad-khalil-imagex</a>) [1], AKHIL BABU (<a href="https://www.drupal.org/u/AkhilBabu">AkhilBabu</a>), Bálint Kléri (<a href="https://www.drupal.org/u/balintbrews">balintbrews</a>), Chandan Singh (<a href="https://www.drupal.org/u/chandu7929">chandu7929</a>), <a href="https://www.drupal.org/u/dpi">dpi</a>, Feliksas Mazeikis (<a href="https://www.drupal.org/u/f.mazeikis">f.mazeikis</a>), Fons Vandamme (<a href="https://www.drupal.org/u/f0ns">f0ns</a>), Georgi Ivanov (<a href="https://www.drupal.org/u/givanov">givanov</a>), John Locke (<a href="https://www.drupal.org/u/freelock">freelock</a>), Jordan Graham (<a href="https://www.drupal.org/u/COBadger">COBadger</a>), julio retkwa (<a href="https://www.drupal.org/u/julioretkwa">julioretkwa</a>), Konstantine Kirkitadze (<a href="https://www.drupal.org/u/kotekirkitadze">kotekirkitadze</a>), Lauri Timmanee (<a href="https://www.drupal.org/u/lauriii">lauriii</a>), Matt Hand (<a href="https://www.drupal.org/u/matthand">matthand</a>), Narendra Singh Rathore (<a href="https://www.drupal.org/u/NarendraR">NarendraR</a>), Ted Bowman (<a href="https://www.drupal.org/u/tedbow">tedbow</a>), utkarsh_33 (<a href="https://www.drupal.org/u/Utkarsh_33">Utkarsh_33</a>), Wim Leers (<a href="https://www.drupal.org/u/wimleers">wimleers</a>), <a href="https://www.drupal.org/u/wotnak">wotnak</a></p> --- <h3>Context Control Center (CCC)</h3> <p>The project is conducting structured quality assurance testing ahead of its first release candidate. Over the past 24 hours, testing activity has uncovered functional issues and identified usability improvements needed before the module can be considered stable.</p> <p>Testing of context item creation workflows has been completed. Separate test coverage is continuing for context scope configuration and context item settings. This phase of testing is validating that the core content management functions—creating, editing, publishing, and organizing context items—work reliably for editors and do not result in data loss or incorrect behavior.</p> <p>A bug was identified where the revision setting is not being respected: <a href="https://git.drupalcode.org/project/ai_context/-/work_items/3586515">Disabling "Create new revisions by default" still creates a new revision</a>. This issue affects content governance and audit trails for organizations that want to control when changes are tracked.</p> <p>Minor interface improvements have been flagged, including simplifying button labels and expanding configuration summaries to show subscription details. Accessibility testing of the dashboard interface has also been assigned and is underway.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize resolution of the revision settings bug to ensure content governance features work correctly before release.</li> <li>Confirm resource allocation for completing accessibility testing and any remediation work that emerges from those audits.</li> </ul> --- <h3>AI CKEditor</h3> <p>The AI CKEditor module enables content editors to use AI assistance directly within the text editor. Over the past 24 hours, work focused on improving how AI understands the context of content being edited.</p> <p>A solution is now in development to address a limitation where AI context was only available for already-saved content. When editors create new content or work with drafts, the AI should be able to understand what type of content is being created and where it will appear on the site. This allows site administrators to configure AI behavior that automatically applies to specific content types or site sections—for example, providing different writing assistance for blog posts versus product pages. The <a href="https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848">Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections</a> fix is currently under review and will make the AI assistance more consistent across both new and existing content.</p> <p>This work supports the module's goal of providing context-aware AI assistance that adapts to different editorial workflows without requiring manual configuration for each editing session.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize review and testing resources to validate the context-handling fix and move it toward completion.</li> <li>Ensure documentation resources are available to help site builders configure context rules once the feature is stable.</li> </ul> <p>Contributors: Kristen Pol (<a href="https://www.drupal.org/u/kepol">kepol</a>)</p> --- <h3>AI Dashboard</h3> <p>The AI Dashboard project saw progress on its ecosystem tracking capability during this period. A question was raised about <a href="https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221">how often ecosystem data is updated</a>, prompting immediate action. In response, the ecosystem data was refreshed and now reflects 234 modules and 38 recipes as of September 10. The update also addressed technical issues that were preventing the system from reliably gathering information about AI-related projects in the Drupal community.</p> <p>This work supports the project's goal of maintaining an accurate view of the AI tooling landscape in Drupal, helping stakeholders understand what solutions are available and how the ecosystem is evolving. Current, reliable ecosystem data enables better decision-making about which AI capabilities to invest in or integrate.</p> <h4>How can I help on this project?</h4> <ul> <li>Clarify governance around ecosystem data refresh frequency: establish whether daily, weekly, or on-demand updates best serve stakeholder needs.</li> <li>Promote the ecosystem dashboard to relevant teams or partners who may benefit from visibility into available AI modules and recipes.</li> </ul> <p>Contributors: Rob Loach (<a href="https://www.drupal.org/u/robloach">robloach</a>) [1]</p> --- <h3>Field Widget Actions</h3> <p>The Field Widget Actions module made meaningful progress toward improved usability and stability over the past 24 hours. Two fixes were merged that address gaps in the configuration experience and system reliability.</p> <p>A new feature now displays a summary of configured actions directly on the form management screen, eliminating the need to open each field's settings individually to see which fields have actions attached. This improves visibility and oversight for teams managing forms with multiple fields.</p> <p>A bug causing system crashes when displaying action descriptions was also resolved. The issue occurred when certain text formatting objects were returned in an incompatible format, leading to failures during rendering.</p> <p>Three open issues remain under active discussion. These address widget compatibility challenges with certain input types, error messaging when required configuration is missing, and a problem preventing users from successfully removing actions once added. All are typical refinements expected during the module's current development phase.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize triage and resolution of the three open usability issues, particularly the removal bug, to reduce friction for teams configuring forms.</li> <li>Ensure adequate testing resources are allocated to validate widget compatibility across diverse field types before the next release.</li> </ul> <p>Contributors: Artem Dmitriiev (<a href="https://www.drupal.org/u/a.dmitriiev">a.dmitriiev</a>), Joonas Meriläinen (<a href="https://www.drupal.org/u/mErilainen">mErilainen</a>), Nik LePage (<a href="https://www.drupal.org/u/niklp">niklp</a>)</p> --- <h3>AI translate</h3> <p>The AI translate module enables automatic translation of Drupal content using AI services. Over the past 24 hours, two significant issues emerged affecting the quality and completeness of translated content.</p> <p>A regression was identified where content built using Layout Builder is no longer being translated. This occurred following recent security improvements. The underlying issue is that layout fields are being silently excluded from translation without warning, meaning users would see apparently successful translations that are actually incomplete. A <a href="https://git.drupalcode.org/project/ai_translate/-/merge_requests/49">fix for the Layout Builder regression</a> has been prepared and is under review.</p> <p>Separately, a <a href="https://git.drupalcode.org/project/ai_translate/-/work_items/3585525">bug was reported</a> where HTML formatting in rich text fields is being corrupted during translation. Formatted content containing bold, italics, links, and other HTML markup is emerging with broken or altered structure, affecting any content type using visual editors like CKEditor. This issue remains open and unaddressed.</p> <p>Both issues impact the reliability of translated content and require resolution before the module can be considered production-ready for sites with formatted or layout-driven content.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize technical review and approval of the pending Layout Builder fix to restore translation completeness</li> <li>Allocate development resources to investigate and resolve the HTML corruption issue, which currently has no proposed solution</li> </ul> <p>Contributors: Sven Decabooter (<a href="https://www.drupal.org/u/svendecabooter">svendecabooter</a>), Carlos Busto Capeáns (<a href="https://www.drupal.org/u/cbccharlie">cbccharlie</a>)</p> --- <h3>ElevenLabs</h3> <p>The ElevenLabs module, which generates audio from text content, is addressing a significant quality issue that affects user experience. A problem was identified where audio playback cuts off after approximately 5,000 characters, even though the service supports twice that length. This means users are only hearing half of the content they expect, which directly impacts accessibility and content delivery.</p> <p>The issue stems from how longer text is processed: content is split into segments, converted to audio separately, and then reassembled. The current method of joining these audio segments is causing truncation. A <a href="https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581">fix for the audio truncation issue</a> has been proposed with a merge request awaiting review. Once resolved, this will restore full-length audio generation and ensure content is delivered as intended.</p> <p>The module remains in active development with normal issue resolution activity for a project at this stage.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize review and approval of the pending fix to restore full audio functionality for end users.</li> <li>Consider allocating resources for broader quality assurance testing to identify similar edge cases before they reach production.</li> </ul> <p>Contributors: Anne Bonham (<a href="https://www.drupal.org/u/annebonham">annebonham</a>)</p> --- <h3>FlowDrop</h3> <p>FlowDrop made progress on stability and configuration control over the past 24 hours. A defect that prevented UI themes from rendering correctly in the editor was resolved. Previously, stylesheets were being inserted into the wrong part of the page, causing visual themes to fail. This issue is now fixed, ensuring the editor displays as intended.</p> <p>A separate fix addressed a problem in the chat feature where conversations involving many tool calls would stop responding when memory limits were reached. This should improve reliability for users working with AI-assisted workflows.</p> <p>New site-level controls were added, allowing administrators to enable or disable the Command Console and AI Assistant features independently. A deprecated "Sync to Cloud" option was also removed from the interface.</p> <p>The underlying JavaScript library was updated to version 2.8.0, and a new alpha release (2.6.0-alpha1) was prepared incorporating these changes. Documentation and tooling configurations were updated accordingly.</p> <p>One <a href="https://git.drupalcode.org/project/flowdrop/-/work_items/3592452">issue</a> remains open related to the page rendering fix, though the underlying work has been completed.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize user testing of the new alpha release to validate the editor rendering and chat reliability fixes before a stable release.</li> <li>Clarify the roadmap for feature toggles: determine which capabilities should remain optional versus becoming defaults.</li> </ul> <p>Contributors: Shibin Das [8]</p> ======================================================================== ## [32] factcheck:tldr:executive ======================================================================== Layer-1 corrections: 0 Judge flags: 0 ======================================================================== ## [33] generate2026Capabilities ======================================================================== You are an editor writing a strategic section for a non-technical executive audience. The Drupal AI initiative has 8 planned capabilities for 2026. Your task is to read the module activity summaries below and explain how today's progress moves each capability forward — or note where there was no relevant progress today. Be concise and direct. Focus on what matters to a business leader: is progress happening, what is the next milestone, and are there any risks? When you describe progress on a capability, add inline citations using [N] where N is the reference number from the list below. You may cite multiple references per item, e.g. [1][3]. Place citations at the end of the relevant sentence, before the closing punctuation. Output an HTML fragment using exactly this structure: <h4>2026 Capabilities Progress</h4> <ol class="capabilities-2026"> <li><strong>Capability name</strong> — One or two sentences on how today's activity relates to this goal, or "No direct progress today" if nothing relevant occurred.</li> </ol> Use only the 8 capabilities listed below, in the same order. Do not add, remove, or rename them. Do not output any text outside the <h4> and <ol> tags. --- 2026 CAPABILITIES --- 1. Page generation — Describe what you need and get a usable page, built from your actual design system components 2. Context management — A central place to define brand voice, style guides, audience profiles, and governance rules that AI can use 3. Background agents — AI that works without being prompted, responding to triggers and schedules while respecting editorial workflows 4. Design system integration — AI that builds with your components and can propose new ones when needed 5. Content creation and discovery — Smarter search, AI-powered optimization, and content drafting assistance 6. Advanced governance — Batch approvals, branch-based versioning, and comprehensive audit trails for AI changes 7. Intelligent website improvements — AI that learns from performance data, proposes concrete changes, and gets smarter over time through editorial review 8. Multi-channel campaigns — Create content for websites, social, email, and automation platforms from a single campaign goal --- REFERENCE LIST --- [1] LLM: Metatag generates and stores tags that were never configured, and keeps only one of them (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586723 [2] Allow Field Widget Actions to target a specific automator 2.x (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586657 [3] Review AI issue process and tagging based on recent usage and update docs (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3550025 [4] Create AI assistant API session only after the first message (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3554797 [5] Add interactive refinement capability to Field Widget Actions 2.x (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586655 [6] Automator FWA buttons give no user feedback when values don't visibly change (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3585987 [7] Restrict access to the AI API Explorer permission (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586733 [8] SetupVdbServer and SetupVdbIndex config actions aren't idempotent, recipe re-apply fails (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586729 [9] Adding Disable Option for AI Automators (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3508800 [10] Bug — the "Content Entity Seeder" tool generates a JSON Schema with one extra array level and breaks when creating/editing content (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586732 [11] textToImage() base64-decodes the image payload twice (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586730 [12] Regression in ai_translate: Layout Builder content is no longer translated (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586731 [13] Streaming chat forces plain-text hostname filtering, corrupting URLs inside HTML attributes (e.g., href="") (ai, Issue) — https://git.drupalcode.org/project/ai/-/work_items/3586624 [14] Issue #3586657: Allow Field Widget Actions to target a specific automator 2.x (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1999 [15] Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x" (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2002 [16] feat: #3585987 by petar.basic: Add status message after AI Field Widget Action automator runs. (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1540 [17] Issue #3586655: "Add interactive refinement capability to Field Widget Actions 2.x" (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1966 [18] Draft: fix: Fix AI Settings form to display AI Platforms and Inference Providers (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/1901 [19] fix: #3586729 Add existence checks to `SetupVdbServer` and `SetupVdbIndex` config actions (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2000 [20] bug: #3586730 textToImage() base64-decodes the image payload twice (ai, MR) — https://git.drupalcode.org/project/ai/-/merge_requests/2001 [21] Support media entities for vision models (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289 [22] Stable 1.0.0 release for use in Drupal CMS recipes (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025 [23] Replace token references by roken replacements (ai_integration_eca, Issue) — https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027 [24] fix: #3585027 Use token replacement for the token input and mask fields (ai_integration_eca, MR) — https://git.drupalcode.org/project/ai_integration_eca/-/merge_requests/26 [25] Add missing OpenAI TTS voices to api_defaults.yml (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3522771 [26] Add test coverage for Whisper (speechToText) and TTS (textToSpeech) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3616894 [27] Support batched embeddings (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568659 [28] Moderation endpoint has hardcoded model (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3607418 [29] "Selected API Key not working" after update to AI 1.3.0 (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3579428 [30] Add reasoning output (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3574226 [31] Change the way that system prompt is being added in 1.2.x (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3573133 [32] Add AI usage checkboxes to ai_provider_openai issue template (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3568620 [33] Add AI 2.x / Search AI compatible OpenAI embedding provider (2.x branch) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3564086 [34] Support File Handling (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3548526 [35] Remove OpenAI SDK dependency and extend it from Drupal AI core module (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531260 [36] Handle pdf files in chat (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3529603 [37] Version of modules can go out of alignment (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519419 [38] Abstract token usage support (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3519302 [39] Usage stats in streaming chat need conversion to an array. (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486224 [40] Log flagged prompts (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3507407 [41] Moderation log overview (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3516044 [42] Streaming chat responses gives Call to undefined method Drupal\ai_provider_openai\OpenAiChatMessageIterator::getTools() (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3515885 [43] Allow skipping of moderations for some embeddings (not all) (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3510601 [44] Bug: Key validation and Rate Limit tester are bypassing host variable; causing Auth errors or Network errors, etc. (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3496412 [45] Selecting "o1" model on automator produces error: "Unsupported parameter: 'max_tokens' is not supported" (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3508042 [46] Error with OpenAI o1 models (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489201 [47] Moderation reads confidence scores instead of the actual violation flags, so almost every category shows as flagged (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161 [48] Error: Call to undefined method OpenAI\Client::responses() in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->chat() (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601 [49] Add support for fast mode (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211 [50] Unable to do GPT 5.4 calls (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729 [51] How to enable type: web_search in Chat? (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3613029 [52] Automated Drupal 12 compatibility fixes for ai_provider_openai 1.3.x-dev (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3606117 [53] Create a skill to setup the provider (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3605812 [54] Automated Drupal 12 compatibility fixes for ai_provider_openai 1.2.x-dev (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3594765 [55] Opt-in to cache embedding lookups (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3552523 [56] Turn off moderation if omni-moderation-latest is not available (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3534855 [57] Add gpt-image-1 image to image capabilitites (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3531289 [58] Warning appears while using automators for image type field "Warning: Undefined array key "b64_json" in Drupal\ai_provider_openai\Plugin\AiProvider\OpenAiProvider->textToImage() " (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3504849 [59] Add event triggering for streaming (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3489846 [60] Improve documentation on setting key (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3486487 [61] Enable Drupal AI Module Functionality for Environments Behind a Proxy (ai_provider_openai, Issue) — https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3481664 [62] Reasoning models return an empty message, and model-specific parameters are rejected by Converse (ai_provider_aws_bedrock, Issue) — https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/work_items/3594715 [63] Issue #3594715: Read the first content block that carries text, and route model-specific parameters to additionalModelRequestFields (ai_provider_aws_bedrock, MR) — https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19 [64] Structured output: an invalid JSON schema is silently accepted instead of throwing an exception (ai_provider_azure, Issue) — https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013 [65] Add batch embeddings (ai_vdb_provider_postgres, Issue) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3610509 [66] indexItems() throws TypeError on array_diff() because $skipDeleteItemIds is never declared (ai_vdb_provider_postgres, Issue) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/work_items/3614504 [67] Issue #3614504: Fix TypeError in indexItems() from undeclared $skipDeleteItemIds property (ai_vdb_provider_postgres, MR) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/33 [68] feat: #3610509 Add batch embeddings (ai_vdb_provider_postgres, MR) — https://git.drupalcode.org/project/ai_vdb_provider_postgres/-/merge_requests/30 [69] Dependency Dashboard (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591657 [70] [>=11.4.7] Broaden our check to determine if a recipe is being applied, use 11.4.7's new `RecipeRunner::isApplying()` (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591985 [71] RecipeAppliedEvent listener installs a module and invalidates the container for every listener after it (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591996 [72] [PP-1] Add logic + docs to not ship Component config entities with Drupal extensions (modules, themes, recipes), and how to do it anyway when there's a good enough reason (pre-populated component trees in Canvas Pages or ContentTemplates) (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3565712 [73] Allow Canvas Patterns+PageRegions+ContentTemplates provided by Recipes to NOT specify component versions (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3571366 [74] Canvas AI: Add Playwright coverage for the chat Tools dropdown (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591951 [75] Config in language other than site default language causes an exception (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591850 [76] Canvas AI: Proper chat history management in the backend (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592032 [77] Canvas AI: Prevent the selected Tool from changing during a chat turn (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591952 [78] Site template installation fails with Canvas 1.11 (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592052 [79] Canvas AI: Unify error handling in `canvas_component_agent`'s tools (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592021 [80] Remove Canvas-only text format guidelines from non-Canvas forms (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592061 [81] <iframe sandbox> + `Window.postMessage()` to add defense in depth (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3527515 [82] Canvas AI: Update orchestrator for content-only / page-variant chrome (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592060 [83] Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592030 [84] Normalize agent-written prop values against their schema and report values the component source cannot store (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3592002 [85] Canvas AI: Unify error handling in the dev page builder agent's tools (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591958 [86] Add component label to settings form (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3591934 [87] The error message about "collapsed input syntax" could be more helpful (canvas, Issue) — https://git.drupalcode.org/project/canvas/-/work_items/3558318 [88] Check for the recipe runner (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1554 [89] refactor(Conflict resolution): #3591785 Strengthen AutoSaveManager to generate stable hashes (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1356 [90] Add docs/recipes.md: how Canvas Component versions interact with recipes... (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1634 [91] Draft: Resolve "Non HTML-validating raw template fails to add Canvas slot wrapper comments" (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1638 [92] #3592031: Added padding to the container to fix visual gaps. (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1600 [93] Draft: 🤖 feat: #3592009 Add locale-aware date/time formatting filters (Twig) and functions (JS) (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1632 [94] 🤖 bug: #3591850 Use entity langcode, not site default, as default translation (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1400 [95] Draft: Resolve "Support mapping multi-valued list fields to array-type component props" (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1630 [96] Improve the Brand Kit Fonts UI (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1520 [97] #3591952 Canvas AI: Prevent the selected Tool from changing during a chat turn (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1604 [98] chore(deps): update dependency cspell to v10 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1273 [99] chore(deps): update dependency @eslint/compat to v2 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1306 [100] chore(deps): update dependency @chromatic-com/storybook to v5 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1305 [101] fix(deps): update npm minor and patch (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1238 [102] chore(deps): update mysql docker tag to v8.3 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1272 [103] bug: #3592061 Remove Canvas-only text format guidelines from non-Canvas forms (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1637 [104] Iframe sandbox: drop allow-same-origin, add postMessage defense in depth (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1610 [105] Draft: task(AI): #3592060 Canvas AI: Update orchestrator for content-only / page-variant chrome (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1636 [106] task(AI): #3592030 Canvas AI: Remove obsolete region-based AI instructions and the unused page variant descriptions form (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1624 [107] 3592021: Add AiResponseValidator::formatErrors() to render accumulated tool... (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1626 [108] issue #3591965: add condition to just load lib on required theme (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1529 [109] test(AI): #3591951 Canvas AI: Add Playwright coverage for the chat Tools dropdown (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1607 [110] feat(AI): #3592032 Canvas AI: Keep the agent history across chat turns in the backend (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1635 [111] Unify error handling in all tools used by the dev page builder agent. (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1584 [112] chore: #3592007 Allow developer and test tooling to locate the Drupal root when the module checkout lives outside the Drupal tree (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1573 [113] Sync brand kit colors through canvas.brand-kit.json in the Canvas CLI (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1586 [114] fix(Data model): #3558318 The error message about "collapsed input syntax" could be more helpful (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1633 [115] Issue #3592055 by f0ns: Hide auto-created folders whose Components are all disabled (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1628 [116] fix(CLI Tool): #3592059 Improve Canvas CLI compatibility with Canvas <1.11 (canvas, MR) — https://git.drupalcode.org/project/canvas/-/merge_requests/1631 [117] CCC Pre-rc1 QA: Context items settings testing (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586496 [118] Disabling "Create new revisions by default" still creates a new revision (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586515 [119] Update context consumer listing configuration summary column (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586514 [120] Change Add AI Context Item button to Add Context Item (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586513 [121] Accessibility Testing: CCC overview dashboard (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586434 [122] CCC Pre-rc1 QA: Test context item creation and workflow (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586491 [123] CCC Pre-rc1 QA: Context scope testing (ai_context, Issue) — https://git.drupalcode.org/project/ai_context/-/work_items/3586492 [124] Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections (ai_ckeditor, Issue) — https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848 [125] Issue #3615848: Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections. (ai_ckeditor, MR) — https://git.drupalcode.org/project/ai_ckeditor/-/merge_requests/16 [126] How often is the ecosystem data updated? (ai_dashboard, Issue) — https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221 [127] Issue #3585221: Rebuild the ecosystem (ai_dashboard, MR) — https://git.drupalcode.org/project/ai_dashboard/-/merge_requests/34 [128] Target element lookup fails for select widgets (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578828 [129] Handle empty "Set widget field value" gracefully (with helpful message) (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578827 [130] Bug when trying to "Remove action"? (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578826 [131] Configured actions are invisible in the "Manage form display" summary row (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578830 [132] TranslatableMarkup returned as string causes fatal error (field_widget_actions, Issue) — https://git.drupalcode.org/project/field_widget_actions/-/work_items/3578829 [133] Add action summary to manage fields ui (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/35 [134] #3578829: Cast return value to string in getDescription() (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/34 [135] Add action summary to manage fields ui (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/32 [136] #3578829: Cast return value to string in getDescription() (field_widget_actions, MR) — https://git.drupalcode.org/project/field_widget_actions/-/merge_requests/33 [137] HTML markup is corrupted/not preserved when translating formatted text fields (text_with_summary, text_long) (ai_translate, Issue) — https://git.drupalcode.org/project/ai_translate/-/work_items/3585525 [138] Regression: Layout Builder content is no longer translated (ai_translate, Issue) — https://git.drupalcode.org/project/ai_translate/-/work_items/3585553 [139] Resolve "Regression: Layout Builder content is no longer translated" (ai_translate, MR) — https://git.drupalcode.org/project/ai_translate/-/merge_requests/49 [140] Text-to-speech audio is truncated to the first ~5,000 characters because MP3 chunks are joined by raw byte concatenation (elevenlabs, Issue) — https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581 [141] Issue #3597581 by banoodle: Fix audio truncated to the first chunk when text... (elevenlabs, MR) — https://git.drupalcode.org/project/elevenlabs/-/merge_requests/5 [142] Bare editor/playground page leaks <head-placeholder>, pushing every stylesheet into <body> so UI themes (Drafter, Minimal) and skins never apply (flowdrop, Issue) — https://git.drupalcode.org/project/flowdrop/-/work_items/3592452 --- TODAY'S MODULE SUMMARIES --- <h3>AI (Artificial Intelligence)</h3> <p>Development activity this period focused on stabilizing Field Widget Actions (FWA) capabilities and addressing defects in several automators and providers. Interactive refinement features for Field Widget Actions were successfully integrated into the 2.x version, giving editors the ability to iteratively improve AI-generated content within forms. A parallel effort to allow FWA buttons to target specific automators rather than running all configured automators simultaneously is now under review.</p> <p>Several bugs were identified and are being corrected. A critical <a href="https://git.drupalcode.org/project/ai/-/work_items/3586730">image generation defect</a> where images were corrupted due to double-decoding was addressed, preventing silent data corruption in generated files. A <a href="https://git.drupalcode.org/project/ai/-/work_items/3586729">configuration idempotency issue</a> affecting recipe-based deployments was also reported and now has a fix under review. Additional bugs surfaced in the Metatag automator (storing all tags rather than only configured ones), streaming chat (corrupting URLs in HTML attributes), and content translation for Layout Builder pages.</p> <p>One <a href="https://git.drupalcode.org/project/ai/-/work_items/3586733">security permission</a> issue has been flagged as a release blocker: the AI API Explorer tool's permission is not marked security-sensitive, despite allowing direct invocation of AI providers and function calls outside normal access controls.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize review and testing of the release blocker permission fix to ensure secure deployment readiness.</li> <li>Allocate resources to address the critical Layout Builder translation regression impacting multilingual content workflows.</li> </ul> <p>Contributors: Abhisek Mazumdar (<a href="https://www.drupal.org/u/abhisekmazumdar">abhisekmazumdar</a>), Artem Dmitriiev (<a href="https://www.drupal.org/u/a.dmitriiev">a.dmitriiev</a>), Dezső Biczó (<a href="https://www.drupal.org/u/mxr576">mxr576</a>), Juan Correa (<a href="https://www.drupal.org/u/jucs7">jucs7</a>), Petar Bašić (<a href="https://www.drupal.org/u/petar_basic">petar_basic</a>), Sven Decabooter (<a href="https://www.drupal.org/u/svendecabooter">svendecabooter</a>)</p> --- <h3>AI Integration - ECA</h3> <p>The AI Integration - ECA module progressed significantly toward production readiness during this period. Three issues were closed, including the completion of <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3534289">Support media entities for vision models</a>, which enables the module to process image and media files for AI-powered categorization and analysis. This expands the module's capability beyond text-only prompts, allowing organizations to automate visual content workflows.</p> <p>A critical stability milestone was achieved with the closure of <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585025">Stable 1.0.0 release for use in Drupal CMS recipes</a>. The module is now production-ready and compatible with Drupal CMS recipe requirements, removing a packaging barrier that prevented adoption by products requiring stable releases. With 934 sites currently using the module, this opens the door to broader distribution.</p> <p>Internal consistency was improved through <a href="https://git.drupalcode.org/project/ai_integration_eca/-/work_items/3585027">Replace token references by token replacements</a>, which standardizes how dynamic content is handled across the module's features, building on patterns established in the media support work.</p> <h4>How can I help on this project?</h4> <ul> <li>Promote the 1.0.0 stable release to downstream teams and partners building Drupal CMS recipes or packaged products.</li> <li>Consider case studies or documentation showcasing vision model use cases to drive adoption of the new media capabilities.</li> </ul> <p>Contributors: Jürgen Haas (<a href="https://www.drupal.org/u/jurgenhaas">jurgenhaas</a>) [1]</p> --- <h3>OpenAI Provider</h3> <p>The OpenAI Provider module saw primarily administrative activity over the last 24 hours, with 23 issues closed. These closures reflect past work being finalized rather than new development landing in the reporting period. The closed issues span multiple areas: support for additional text-to-speech voices, batched embeddings for faster vector database indexing, reasoning output from AI models, file handling capabilities, moderation logging interfaces, and compatibility fixes for various OpenAI model parameters. This cleanup work brings the project's issue queue into better alignment with its current functionality.</p> <p>Several active issues remain open and require attention. A <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3622161">moderation accuracy bug</a> is causing almost every content category to appear flagged because the system reads confidence scores instead of actual violation flags. An <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3620601">integration error</a> prevents all chat operations when using version 1.3.0-beta2 with the stable AI module. Two feature requests relate to supporting OpenAI's newer capabilities: <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3619211">fast mode</a> for improved response times and <a href="https://git.drupalcode.org/project/ai_provider_openai/-/work_items/3614729">GPT 5.4 compatibility</a>.</p> <p>The project has 42 open issues in total, including work underway to improve test coverage for speech-to-text and text-to-speech operations. This is expected and appropriate for a module in active development supporting a rapidly evolving external API.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize the moderation accuracy bug and chat integration error, as these block core functionality for users on current stable versions.</li> <li>Clarify the module's roadmap for supporting new OpenAI features like fast mode and GPT 5.4, particularly whether these require upstream AI framework changes.</li> </ul> --- <h3>AWS Bedrock Provider</h3> <p>A bug affecting reasoning models was identified and a fix has been prepared. When using reasoning models, the system was returning empty responses instead of the actual generated text. This occurred because reasoning models structure their output differently, placing reasoning content in a separate block before the main response. The current code was looking in the wrong location for the text.</p> <p>The same issue affects the handling of model-specific configuration parameters, which were being rejected instead of properly routed to the underlying service.</p> <p>A <a href="https://git.drupalcode.org/project/ai_provider_aws_bedrock/-/merge_requests/19">fix for both problems</a> is now under review. Once merged, reasoning models will work correctly and model-specific parameters will be properly supported. This expands the module's compatibility with advanced AI capabilities and gives users more control over model behavior.</p> <h4>How can I help on this project?</h4> <ul> <li>Ensure testing resources are available to validate the reasoning model fix before release</li> <li>Consider prioritizing review capacity to move the pending fix through the approval process</li> </ul> Contributors: Besart Bytyqi (<a href="https://www.drupal.org/u/besart">besart</a>) --- <h3>Microsoft Azure AI</h3> <p>The Microsoft Azure AI provider module saw issue reporting activity but no code changes during this period. A <a href="https://git.drupalcode.org/project/ai_provider_azure/-/work_items/3618013">validation gap in structured output handling</a> was documented, where invalid JSON schemas are silently accepted rather than flagged. This means that configuration errors—such as marking fields as required without defining them—are not caught early, potentially leading to unexpected behavior downstream. The issue is open and unassigned, with no fix currently in development.</p> <p>This represents normal maintenance activity for a module in active development. The issue has been logged and is awaiting prioritization and assignment.</p> <h4>How can I help on this project?</h4> <ul> <li>Allocate developer time to triage and assign the open validation issue, ensuring it is prioritized appropriately within the roadmap.</li> <li>Consider resourcing a quality assurance review of schema validation across the module to identify similar gaps proactively.</li> </ul> --- <h3>Postgres VDB Provider</h3> <p>The Postgres VDB Provider is being adapted to support batch embeddings, a new capability that enables indexing systems to process and store multiple content chunks in a single request instead of one at a time. This change is expected to meaningfully reduce indexing time for sites using this provider. The work aligns the Postgres provider with improvements already made to other vector database providers in the AI module ecosystem. A merge request implementing this capability is now open for review.</p> <p>Separately, a bug was identified and addressed that prevented any content indexing from working at all. When attempting to index content through Search API, the system would fail with an error due to an uninitialized variable. A fix has been prepared and is awaiting merge.</p> <p>These updates move the module toward more efficient operation and restore core indexing functionality that was broken.</p> <h4>How can I help on this project?</h4> <ul> <li>Ensure reviewer capacity is available to evaluate and merge the two open merge requests so the performance improvements and bug fix reach users promptly.</li> <li>Consider prioritizing resources for testing the batch embeddings feature in realistic production scenarios to validate the performance gains.</li> </ul> Contributors: Sergiu Nagailic (<a href="https://www.drupal.org/u/Nikro">Nikro</a>), Zeeshaan khann (<a href="https://www.drupal.org/u/zeeshan_khan">zeeshan_khan</a>) --- <h3>Drupal Canvas</h3> <h4>What shipped</h4> <p>A significant fix to the AI chat experience was completed: the <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591952">selected Tool can no longer change mid-turn</a>. Previously, because each chat turn involved multiple backend requests, users could inadvertently switch or clear the AI tool while a conversation was still running, causing inconsistent behavior. The system now locks the tool selection for the duration of a turn, both in the interface and on the server.</p> <h4>In progress</h4> <p>Work continues on several fronts. An <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591996">installation failure affecting site templates</a> was identified and is under review—the issue causes a container error when applying recipes with Canvas 1.11. Separately, efforts are underway to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592030">remove outdated AI instructions tied to the old region-based layout model</a>, now replaced by page variants. Progress is also being made on <a href="https://git.drupalcode.org/project/canvas/-/work_items/3591958">unifying error handling across AI agent tools</a> and improving <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592002">validation of AI-generated content</a> to reduce unnecessary rejections. On the developer experience side, work is moving forward to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592007">support module development outside the Drupal tree</a> and to <a href="https://git.drupalcode.org/project/canvas/-/work_items/3592019">sync brand kit colors through the Canvas CLI</a>.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize diagnosis and resolution of the site template installation failure, which may block adoption of Canvas 1.11 in production environments.</li> <li>Ensure developer tooling improvements receive adequate review and testing resources to accelerate contributor onboarding.</li> </ul> <p>Contributors: **** (<a href="https://www.drupal.org/u/project_19391_bot_cb443e60680723bba08a5351bf595251">project_19391_bot_cb443e60680723bba08a5351bf595251</a>), Adam G-H (<a href="https://www.drupal.org/u/phenaproxima">phenaproxima</a>), Ahmad Khalil (<a href="https://www.drupal.org/u/ahmad-khalil-imagex">ahmad-khalil-imagex</a>) [1], AKHIL BABU (<a href="https://www.drupal.org/u/AkhilBabu">AkhilBabu</a>), Bálint Kléri (<a href="https://www.drupal.org/u/balintbrews">balintbrews</a>), Chandan Singh (<a href="https://www.drupal.org/u/chandu7929">chandu7929</a>), <a href="https://www.drupal.org/u/dpi">dpi</a>, Feliksas Mazeikis (<a href="https://www.drupal.org/u/f.mazeikis">f.mazeikis</a>), Fons Vandamme (<a href="https://www.drupal.org/u/f0ns">f0ns</a>), Georgi Ivanov (<a href="https://www.drupal.org/u/givanov">givanov</a>), John Locke (<a href="https://www.drupal.org/u/freelock">freelock</a>), Jordan Graham (<a href="https://www.drupal.org/u/COBadger">COBadger</a>), julio retkwa (<a href="https://www.drupal.org/u/julioretkwa">julioretkwa</a>), Konstantine Kirkitadze (<a href="https://www.drupal.org/u/kotekirkitadze">kotekirkitadze</a>), Lauri Timmanee (<a href="https://www.drupal.org/u/lauriii">lauriii</a>), Matt Hand (<a href="https://www.drupal.org/u/matthand">matthand</a>), Narendra Singh Rathore (<a href="https://www.drupal.org/u/NarendraR">NarendraR</a>), Ted Bowman (<a href="https://www.drupal.org/u/tedbow">tedbow</a>), utkarsh_33 (<a href="https://www.drupal.org/u/Utkarsh_33">Utkarsh_33</a>), Wim Leers (<a href="https://www.drupal.org/u/wimleers">wimleers</a>), <a href="https://www.drupal.org/u/wotnak">wotnak</a></p> --- <h3>Context Control Center (CCC)</h3> <p>The project is conducting structured quality assurance testing ahead of its first release candidate. Over the past 24 hours, testing activity has uncovered functional issues and identified usability improvements needed before the module can be considered stable.</p> <p>Testing of context item creation workflows has been completed. Separate test coverage is continuing for context scope configuration and context item settings. This phase of testing is validating that the core content management functions—creating, editing, publishing, and organizing context items—work reliably for editors and do not result in data loss or incorrect behavior.</p> <p>A bug was identified where the revision setting is not being respected: <a href="https://git.drupalcode.org/project/ai_context/-/work_items/3586515">Disabling "Create new revisions by default" still creates a new revision</a>. This issue affects content governance and audit trails for organizations that want to control when changes are tracked.</p> <p>Minor interface improvements have been flagged, including simplifying button labels and expanding configuration summaries to show subscription details. Accessibility testing of the dashboard interface has also been assigned and is underway.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize resolution of the revision settings bug to ensure content governance features work correctly before release.</li> <li>Confirm resource allocation for completing accessibility testing and any remediation work that emerges from those audits.</li> </ul> --- <h3>AI CKEditor</h3> <p>The AI CKEditor module enables content editors to use AI assistance directly within the text editor. Over the past 24 hours, work focused on improving how AI understands the context of content being edited.</p> <p>A solution is now in development to address a limitation where AI context was only available for already-saved content. When editors create new content or work with drafts, the AI should be able to understand what type of content is being created and where it will appear on the site. This allows site administrators to configure AI behavior that automatically applies to specific content types or site sections—for example, providing different writing assistance for blog posts versus product pages. The <a href="https://git.drupalcode.org/project/ai_ckeditor/-/work_items/3615848">Send bundle and page path in entity_context so AI Context can scope unsaved entities and Site Sections</a> fix is currently under review and will make the AI assistance more consistent across both new and existing content.</p> <p>This work supports the module's goal of providing context-aware AI assistance that adapts to different editorial workflows without requiring manual configuration for each editing session.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize review and testing resources to validate the context-handling fix and move it toward completion.</li> <li>Ensure documentation resources are available to help site builders configure context rules once the feature is stable.</li> </ul> <p>Contributors: Kristen Pol (<a href="https://www.drupal.org/u/kepol">kepol</a>)</p> --- <h3>AI Dashboard</h3> <p>The AI Dashboard project saw progress on its ecosystem tracking capability during this period. A question was raised about <a href="https://git.drupalcode.org/project/ai_dashboard/-/work_items/3585221">how often ecosystem data is updated</a>, prompting immediate action. In response, the ecosystem data was refreshed and now reflects 234 modules and 38 recipes as of September 10. The update also addressed technical issues that were preventing the system from reliably gathering information about AI-related projects in the Drupal community.</p> <p>This work supports the project's goal of maintaining an accurate view of the AI tooling landscape in Drupal, helping stakeholders understand what solutions are available and how the ecosystem is evolving. Current, reliable ecosystem data enables better decision-making about which AI capabilities to invest in or integrate.</p> <h4>How can I help on this project?</h4> <ul> <li>Clarify governance around ecosystem data refresh frequency: establish whether daily, weekly, or on-demand updates best serve stakeholder needs.</li> <li>Promote the ecosystem dashboard to relevant teams or partners who may benefit from visibility into available AI modules and recipes.</li> </ul> <p>Contributors: Rob Loach (<a href="https://www.drupal.org/u/robloach">robloach</a>) [1]</p> --- <h3>Field Widget Actions</h3> <p>The Field Widget Actions module made meaningful progress toward improved usability and stability over the past 24 hours. Two fixes were merged that address gaps in the configuration experience and system reliability.</p> <p>A new feature now displays a summary of configured actions directly on the form management screen, eliminating the need to open each field's settings individually to see which fields have actions attached. This improves visibility and oversight for teams managing forms with multiple fields.</p> <p>A bug causing system crashes when displaying action descriptions was also resolved. The issue occurred when certain text formatting objects were returned in an incompatible format, leading to failures during rendering.</p> <p>Three open issues remain under active discussion. These address widget compatibility challenges with certain input types, error messaging when required configuration is missing, and a problem preventing users from successfully removing actions once added. All are typical refinements expected during the module's current development phase.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize triage and resolution of the three open usability issues, particularly the removal bug, to reduce friction for teams configuring forms.</li> <li>Ensure adequate testing resources are allocated to validate widget compatibility across diverse field types before the next release.</li> </ul> <p>Contributors: Artem Dmitriiev (<a href="https://www.drupal.org/u/a.dmitriiev">a.dmitriiev</a>), Joonas Meriläinen (<a href="https://www.drupal.org/u/mErilainen">mErilainen</a>), Nik LePage (<a href="https://www.drupal.org/u/niklp">niklp</a>)</p> --- <h3>AI translate</h3> <p>The AI translate module enables automatic translation of Drupal content using AI services. Over the past 24 hours, two significant issues emerged affecting the quality and completeness of translated content.</p> <p>A regression was identified where content built using Layout Builder is no longer being translated. This occurred following recent security improvements. The underlying issue is that layout fields are being silently excluded from translation without warning, meaning users would see apparently successful translations that are actually incomplete. A <a href="https://git.drupalcode.org/project/ai_translate/-/merge_requests/49">fix for the Layout Builder regression</a> has been prepared and is under review.</p> <p>Separately, a <a href="https://git.drupalcode.org/project/ai_translate/-/work_items/3585525">bug was reported</a> where HTML formatting in rich text fields is being corrupted during translation. Formatted content containing bold, italics, links, and other HTML markup is emerging with broken or altered structure, affecting any content type using visual editors like CKEditor. This issue remains open and unaddressed.</p> <p>Both issues impact the reliability of translated content and require resolution before the module can be considered production-ready for sites with formatted or layout-driven content.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize technical review and approval of the pending Layout Builder fix to restore translation completeness</li> <li>Allocate development resources to investigate and resolve the HTML corruption issue, which currently has no proposed solution</li> </ul> <p>Contributors: Sven Decabooter (<a href="https://www.drupal.org/u/svendecabooter">svendecabooter</a>), Carlos Busto Capeáns (<a href="https://www.drupal.org/u/cbccharlie">cbccharlie</a>)</p> --- <h3>ElevenLabs</h3> <p>The ElevenLabs module, which generates audio from text content, is addressing a significant quality issue that affects user experience. A problem was identified where audio playback cuts off after approximately 5,000 characters, even though the service supports twice that length. This means users are only hearing half of the content they expect, which directly impacts accessibility and content delivery.</p> <p>The issue stems from how longer text is processed: content is split into segments, converted to audio separately, and then reassembled. The current method of joining these audio segments is causing truncation. A <a href="https://git.drupalcode.org/project/elevenlabs/-/work_items/3597581">fix for the audio truncation issue</a> has been proposed with a merge request awaiting review. Once resolved, this will restore full-length audio generation and ensure content is delivered as intended.</p> <p>The module remains in active development with normal issue resolution activity for a project at this stage.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize review and approval of the pending fix to restore full audio functionality for end users.</li> <li>Consider allocating resources for broader quality assurance testing to identify similar edge cases before they reach production.</li> </ul> <p>Contributors: Anne Bonham (<a href="https://www.drupal.org/u/annebonham">annebonham</a>)</p> --- <h3>FlowDrop</h3> <p>FlowDrop made progress on stability and configuration control over the past 24 hours. A defect that prevented UI themes from rendering correctly in the editor was resolved. Previously, stylesheets were being inserted into the wrong part of the page, causing visual themes to fail. This issue is now fixed, ensuring the editor displays as intended.</p> <p>A separate fix addressed a problem in the chat feature where conversations involving many tool calls would stop responding when memory limits were reached. This should improve reliability for users working with AI-assisted workflows.</p> <p>New site-level controls were added, allowing administrators to enable or disable the Command Console and AI Assistant features independently. A deprecated "Sync to Cloud" option was also removed from the interface.</p> <p>The underlying JavaScript library was updated to version 2.8.0, and a new alpha release (2.6.0-alpha1) was prepared incorporating these changes. Documentation and tooling configurations were updated accordingly.</p> <p>One <a href="https://git.drupalcode.org/project/flowdrop/-/work_items/3592452">issue</a> remains open related to the page rendering fix, though the underlying work has been completed.</p> <h4>How can I help on this project?</h4> <ul> <li>Prioritize user testing of the new alpha release to validate the editor rendering and chat reliability fixes before a stable release.</li> <li>Clarify the roadmap for feature toggles: determine which capabilities should remain optional versus becoming defaults.</li> </ul> <p>Contributors: Shibin Das [8]</p> ======================================================================== ## [34] factcheck:capabilities:executive ======================================================================== Layer-1 corrections: 0 Judge flags: 5