Daily Digest

Daily Digest – 6 August 2026

2026 Capabilities Progress

  1. Page generation — Significant progress: AI-driven page building now sends resolved component property values to enable better understanding of both content and layout [65][89], the component agent supports browser-driven iteration for incremental progress [71], and a region validation fix prevents the system from discarding builds when the AI names non-existent regions [58][94].
  2. Context management — The Context Control Center hardened access controls and documented its trust model [102][134], separated form guidance from scope plugin descriptions for clearer UI [129][135], and opened work on documentation of scope selection and ranking logic [98][132]. The module is progressing toward release candidate.
  3. Background agents — The Views Agent issues were closed with draft or non-merged code [15][43], enabling natural-language configuration of data displays. Field Widget Action targeting was fixed and backported for the main branch to prevent unintended automator triggering [12][14][27][28], improving automation reliability.
  4. Design system integration — Canvas opened work on a settings form to select available AI tools for the chat interface [68][70][92], closed work on component detail retrieval capabilities [69], and merged fixes to improve component property handling during AI-assisted page building [65][89].
  5. Content creation and discovery — Search precision improved with a fix ensuring reranking prefers chunk text over configured indexed fields [4][21]. The Suggest Tags automator now works with the Tagify widget [11][23], enhancing content tagging workflows.
  6. Advanced governance — No direct progress today.
  7. Intelligent website improvements — No direct progress today.
  8. Multi-channel campaigns — No direct progress today.

Shipped

  1. AI module Field Widget Action targeting fix merged and backported — A bug was resolved where field-level automation buttons would trigger all configured automators instead of just the selected one, ensuring more predictable behavior across multiple supported versions [24][26][27][28].
  2. AI Views Agent work closed — Work on a natural-language interface for configuring Views has concluded, though implementation merge requests were closed without merging [15].
  3. AI Agents views capability closed — Work on a new agent type enabling AI-powered creation and management of views has concluded, though associated merge requests were closed without merging [43].
  4. AI Agents tool deletion bug fix in progress — A fix allowing users to delete the last tool from an agent without encountering errors is under review [38][45].
  5. Canvas AI configuration and iteration improvements partially shipped — The page builder now sends resolved property values to the AI agent in requests, with additional work on AI tools selection interface and browser-driven iteration workflow under development [65][70][71][89].

Ongoing

  1. AI module Symfony compatibility issues identified — Runtime type mismatches have been documented that require resolution to maintain compatibility with Symfony AI Platform versions 0.8 through 0.12 [2].
  2. AI Agents cache invalidation issue under development — Newly created agents require manual cache clearing before becoming usable, affecting user experience during agent creation [40][44].
  3. Canvas data integrity and caching bugs reported — Four new issues document problems with component version hashing, cache collisions in form widgets, missing cache tags for array properties, and upgrade path gaps affecting older installations [53][55][56][59].
  4. Context Control Center approaching release candidate — Access control hardening, scope selection documentation, and form rendering improvements are under review as the module prepares for its first stable release [98][102][132][133][134].
  5. AG-UI configuration import fix in review — A fix preventing configuration import issues when using AI agents dependency is awaiting merge [141][142].

AI (Artificial Intelligence)

The AI module made significant progress this period, closing two issues and merging fixes that improve precision in how automated tasks are triggered. The Field Widget Action targeting fix was merged and backported across multiple supported versions, resolving a bug where field-level automation buttons would inadvertently trigger all configured automators instead of just the selected one. This ensures more predictable behaviour when sites run multiple automations on the same field. The Drupal AI Views Agent meta-issue was also closed, marking completion of a natural-language interface for configuring Views.

Work is underway on several features designed to support agent-driven site configuration: skills for managing general AI settings, guardrails, and advanced tagging with the Tagify widget are all progressing through open merge requests. Bug fixes are in the pipeline for issues affecting automator activation, search result reranking at the chunk level, tool-call validation error handling, and test provider type mismatches. Separately, an issue tracking the adoption of changes from Symfony AI Platform versions 0.8 through 0.12 has been opened, identifying runtime type mismatches that will need resolution to maintain compatibility.

Several strategic improvements are under discussion, including support for provider-native tool calling and a modern chatbot frontend library to replace the current tightly coupled interface. Work to block accidental deletion of inference providers that are still referenced by configurations is also planned for the next major version.

How can I help on this project?

  • Prioritise and assign development resources to resolve the Symfony AI Platform compatibility issues, which currently cause runtime failures.
  • Sponsor user experience research and prototyping for the modern chatbot frontend to clarify requirements before implementation begins.

Contributors: Abhisek Mazumdar (abhisekmazumdar) [1], Ann Mary Sruthy (annmarysruthy), Artem Dmitriiev (a.dmitriiev), Bruno Bruno (bbruno), Scott Euser (scotteuser)

View project on GitLab: ai

Drupal AI Initiative

The Drupal AI Initiative continued its partnership expansion work during this period. The onboarding process for Oomph as a new Silver partner progressed, with several administrative tasks completed including mailing list additions and Slack channel access. Follow-up work has been planned to ensure successful integration through Run AI partner audit report: Oomph and to add the partner to public-facing pages via Add Silver AI partner to d.o pages: Oomph.

Two template issues for standardizing future partner onboarding workflows were closed, indicating the establishment of repeatable processes for bringing new partners into the initiative. This standardization should streamline future partnership activations and reduce administrative overhead.

The work reflects ongoing ecosystem development for the AI Initiative, building the partner network that will support the project's long-term sustainability and market reach.

How can I help on this project?

  • Identify and introduce additional potential Silver or Gold partners to accelerate ecosystem growth and funding.
  • Ensure marketing and communications resources are available to support partner visibility once Oomph's public page materials are ready.

View project on GitLab: ai_initiative

AI Integration - ECA

This module connects Drupal's AI capabilities with the Event-Condition-Action (ECA) automation framework. Over the past 24 hours, work has focused on two areas that will make the module easier to configure and more capable.

A new task was opened to simplify structured output configuration in the Chat action. Currently, users must enter complex schema definitions as plain text, which is error-prone and difficult to verify. The proposed improvement would replace this with a purpose-built schema editor already available in the core AI module, reducing configuration errors and improving usability for site builders.

Separately, a draft integration is under review that will allow the module to inspect and modify AI requests before they are sent to providers. This capability unlocks advanced use cases such as cost tracking, content filtering, and auditing. Review feedback was addressed during the reporting period, focusing on ensuring the integration handles metadata consistently when multiple automation rules modify the same request.

How can I help on this project?

Consider prioritizing user testing for the schema editor improvement once available, as configuration usability is often a barrier to adoption. If the interceptor functionality aligns with organizational compliance or cost-control needs, signal that priority to help guide the roadmap toward a stable release.

Contributors: Marcus Johansson (Marcus_Johansson)

View project on GitLab: ai_integration_eca

AI Agents

Progress this period focused on expanding the module's capabilities and resolving bugs that affect usability. Two significant pieces of work were completed: Add views agent was closed, delivering a new agent type that enables AI-powered creation and management of views—a feature that makes it easier for users to generate data displays through natural language interaction. Additionally, a bug preventing users from deleting the last tool from an agent was resolved with the closure of If only one tool, cannot delete it, gives an error.

Several active development threads remain open. Work is underway to address caching issues that prevent newly created agents from being immediately usable until caches are cleared. A minor warning related to missing parameters in tool definitions is being addressed through an open merge request. Two feature requests remain in progress: adding short-term memory capabilities to agents and enabling parent agents to forward requests without modification, both aimed at improving agent efficiency and reducing costs.

How can I help on this project?

  • Prioritize or assign resources to resolve the cache invalidation issue, which currently requires manual intervention and affects the user experience when creating new agents.
  • Consider allocating dedicated testing resources to validate new agent types and memory features as they near completion.

Contributors: Alexis Martínez (alexismmd), Bharat Kelotra (bharatkelotra), Jibran Ijaz (jibran)

View project on GitLab: ai_agents

Tool API

The Tool API project saw continued documentation and developer experience work during this period. Two preparation efforts are underway to make it easier for developers to build plugins and for potential users to understand the module's capabilities.

Work continued on Create a skill for generating Tool plugins (best practices, how-to, placement), which aims to help developers correctly author Tool plugins by guiding them through the required components—such as plugin attributes, typed input and output definitions, and operation classifications. This should reduce the learning curve for teams wanting to expose well-defined actions to AI agents.

Separately, progress was made on Write documentation and enable project pages, which will provide auto-generated project documentation on every main branch update, improving discoverability and onboarding for potential adopters.

Both items have open merge requests in development. No code was merged during the reporting window.

How can I help on this project?

  • Prioritize review and feedback on the two open documentation-focused merge requests to accelerate their completion and improve adoption readiness.
  • Consider allocating testing or user-feedback resources to validate that the developer guidance materials meet real-world onboarding needs.

Contributors: Marcus Johansson (Marcus_Johansson)

View project on GitLab: tool

Drupal Canvas

Over the past 24 hours, the project merged six updates spanning AI capabilities, developer tools, testing infrastructure, and data integrity.

Several fixes were integrated to improve the AI-assisted page builder experience. The system now sends resolved component property values alongside the page structure in a single request parameter, enabling the AI agent to better understand both content and layout when generating pages. Additionally, the AI workflow was restructured to support browser-driven iteration: the agent pauses after each decision and returns control to the client, allowing incremental progress rather than blocking on a single server request. A configuration interface was added to allow administrators to select which AI tools are available in the developer chat interface.

Developer tooling received two fixes. The command-line tool now correctly assigns file extensions (.jsx or .tsx) based on the source code pulled from the server, preventing compilation failures. Dependency packages across the JavaScript stack were updated to current versions.

Test suite efficiency improved: end-to-end tests now distribute across parallel runners based on runtime characteristics rather than file count, reducing overall pipeline duration. Headless preview rendering was extended to support content templates in addition to pages.

Four new issues were opened documenting bugs in data handling and caching. These include problems with component version hashing during single-process installs, cache collisions in dynamically generated form widgets, missing cache tags for array-based component properties, and a data migration gap affecting sites upgrading from earlier releases. Work is underway on fixes for the upgrade path and region validation issues.

How can I help on this project?

  • Ensure the team has sufficient headcount to address the four newly opened data integrity and caching issues alongside ongoing feature development.
  • Support prioritisation decisions around stabilising the upgrade path and editor preview viewport improvements ahead of the next release cycle.

Contributors: **** (project_19391_bot_17a61b0edeafd5a3db3e8b2462b6c010), **** (project_19391_bot_cb443e60680723bba08a5351bf595251), Adam G-H (phenaproxima), Ben Mullins (bnjmnm), Bálint Kléri (balintbrews) [1], Chandan Singh (chandu7929), Christian López Espínola (penyaskito) [1], Dave Long (longwave), David Bravo (davidbp), Feliksas Mazeikis (f.mazeikis), Harumi Jang (hooroomoo), Juan Correa (jucs7) [1], Matt Glaman (mglaman), Rajab Natshah (RajabNatshah), Rajendar Pothireddypally (RajendarReddy), wotnak [2]

View project on GitLab: canvas

Context Control Center (CCC)

The project completed three important improvements during the reporting period as part of preparing for its upcoming release candidate. Work focused on access control, user interface clarity, and code quality.

A comprehensive update to access checks and documentation was merged, ensuring context items are protected correctly and clarifying the trust model for site builders who create and manage content for AI features. This work closes small disclosure gaps and adds test coverage for different permission scenarios.

The team also separated form guidance from scope plugin descriptions, allowing the interface to show context-appropriate help text. For example, warnings relevant to the context item editor no longer appear on unrelated settings pages. This improves clarity without changing functionality.

A final code consistency cleanup was completed, improving maintainability through clearer naming conventions and modern code patterns. Planning and triage work for the current sprint was closed, and two new merge requests addressing scope selection documentation and a form rendering issue were opened for review.

The module is in beta and progressing toward its first stable release. Open issues and ongoing refinements are typical for this stage.

How can I help on this project?

  • Recruit additional beta testers from partner organizations or user groups to validate real-world workflows before the release candidate.
  • Confirm QA team capacity is sufficient to complete testing coverage before the stable release milestone.

Contributors: Juan Correa (jucs7) [1], Kristen Pol (kepol) [1], Tamas Balog (tbalog), Tekla Aivazashvili (aivazashvilit) [1]

View project on GitLab: ai_context

AI Dashboard

The AI Dashboard project completed a significant update to its AI Ecosystem component during this period. The work involved refreshing the module to incorporate recent changes from upstream sources and ensuring that category-based organization is properly implemented. This enhancement will help users better navigate and understand the landscape of AI-related modules available in the Drupal ecosystem. The Update the AI Ecosystem issue was closed after the corresponding merge request was integrated.

This update positions the dashboard to provide more structured and current information about available AI tools, which supports better decision-making for organizations evaluating AI capabilities within Drupal.

How can I help on this project?

  • Review the updated AI Ecosystem categorization and provide feedback on whether it meets strategic needs for evaluating AI module adoption.
  • Identify and communicate priority use cases that would benefit from dashboard visibility or analytics enhancements.
Contributors: Rob Loach (robloach) [1]

View project on GitLab: ai_dashboard

AI translate

The AI Translate module enables automated content translation using AI providers. Over the past 24 hours, a defect was reported that prevents the module from translating media entities such as images. When users attempt to translate media content into a new language, the translation process fails with an error. This issue has been flagged as a quick win, suggesting it may be straightforward to resolve once development work begins.

No code changes were merged or committed during this period. The InvalidArgumentException: Invalid translation language (und) when translating media entities issue remains open and unassigned. As the module is in active development as part of the AI Initiative Sprint, open issues of this nature are typical and expected at this stage.

How can I help on this project?

Consider allocating development resources to address the open media translation defect, which has been identified as achievable in the short term. Support prioritization decisions around which translation capabilities are most valuable to end users and content teams.

View project on GitLab: ai_translate

AG-UI

A configuration issue that could prevent successful imports when using the AI agents dependency has been identified and a fix has been prepared. This work ensures smoother deployment and configuration management workflows, reducing potential disruptions when teams update or synchronize site configurations. The fix is currently under review.

No releases or merges occurred during this period. Development activity remains focused on resolving configuration reliability issues that affect deployment processes.

How can I help on this project?

  • Allocate reviewer capacity to assess and merge the pending configuration fix, which will improve deployment stability.
  • Consider whether additional testing resources are needed to validate configuration workflows before the next release.

Contributors: Takaya Hirose (takayahirose)

View project on GitLab: agui

References

[1] AiAutomatorEntityModifier::entityHasConfig() ignores the AI Automator's disabled (status: false) setting (ai, Issue)

[2] Adopt Symfony AI Platform 0.8-0.12 changes (ai, Issue)

[4] `AiReranker::extractItemText()` should prefer chunk text over configured indexed fields (ai, Issue)

[11] Suggest Tags automator should work with Tagify (ai, Issue)

[12] Allow Field Widget Actions to target a specific automator (ai, Issue)

[14] Allow Field Widget Actions to target a specific automator 2.x (ai, Issue)

[15] [META] Drupal AI Views Agent (ai, Issue)

[21] bug: #3586637 should prefer chunk text over configured indexed fields (ai, MR)

[23] feat: #3536912 Suggest Tags automator should work with Tagify (ai, MR)

[24] Issue #3571915: cherry pick 3e3b5970 for 1.3.x (ai, MR)

[26] Issue #3571915: Cherry-pick for 1.4.x (ai, MR)

[27] bug: #3571915 Filter Field Widget Action automator processing to the selected automator ID (ai, MR)

[28] bug: #3571915 Filter Field Widget Action automator processing to the selected automator ID (ai, MR)

[38] If only one tool, cannot delete it, gives an error: Undefined array key 1 in Drupal\ai\Element\ToolsLibrary::processToolsLibrary() (line 89 of modules/contrib/ai/src/Element/ToolsLibrary.php) & more (ai_agents, Issue)

[40] Connecting an Agent to an Assistant doesn't work until cache clear and Tool API (ai_agents, Issue)

[43] Add views agent (ai_agents, Issue)

[44] test: #3586021 Add regression test for ai_agent plugin cache invalidation (ai_agents, MR)

[45] Issue#3563628:Crash-Error-When-Trying-Remove-Last-Tool-From-Agent fixes (ai_agents, MR)

[53] StaticPropSource widget building is order-dependent: dangling prop fields collide in core's TypedData item-prototype cache (canvas, Issue)

[55] Array prop shapes don't inherit their item shape's cache tags, so prop-shape invalidation misses them (canvas, Issue)

[56] Component version hashes go stale on single-process installs: prop-shape re-resolution is deferred to teardown and skipped on a cold cache (canvas, Issue)

[58] Canvas AI: page builder discards the whole build when the LLM names a region that does not exist (canvas, Issue)

[59] Missing upgrade path: stored `default-relative-url` prop sources written before 1.8 contain `$id` and fail `DefaultRelativeUrlPropSource::parse()` (canvas, Issue)

[65] Canvas AI: Send resolved prop values with current_layout, on both request paths (canvas, Issue)

[68] Canvas AI: Add Tools dropdown to chat UI (canvas, Issue)

[69] Canvas AI: Add the get_component_details AI tool + component-context helper (+ tests) (canvas, Issue)

[70] Canvas AI: Add an Agents & Tools settings form to select the main agent and chat Tools (canvas, Issue)

[71] Canvas AI: Run the component agent in the dev chat with setLooped(FALSE), driven by the browser (canvas, Issue)

[89] #3591883 Canvas AI: Send resolved prop values with current_layout, on both request paths (canvas, MR)

[92] Draft: Resolve "Canvas AI: Add Tools dropdown to chat UI" (canvas, MR)

[94] fix: #3591876 Validate the target region before placing components (canvas, MR)

[98] Review and document context selection ranking factors before 1.0 (ai_context, Issue)

[102] Harden CCC context item access checks and document the authoring trust model (ai_context, Issue)

[129] Decouple context item form descriptions from scope plugin descriptions (ai_context, Issue)

[132] Issue #3586365: Document scope selection and ranking logic. (ai_context, MR)

[133] Issue #3586369: "Global context" checkbox description renders a literal <br> instead of a line break (ai_context, MR)

[134] Issue #3586356: Harden CCC context item access checks and document the authoring trust model. (ai_context, MR)

[135] Issue #3586361: Decouple context item form descriptions from scope plugin descriptions. (ai_context, MR)

[141] Fix ai_agents dependency setting that currently can cause config import issue (agui, Issue)

[142] 3590048: Fix ai_agents dependency setting (agui, MR)

Period: 6 August 2026 to 7 August 2026  |  Generated: 7 August 2026 05:00 GMT

To view the full AI prompts used to generate this digest check this file: 24h_2026-08-06_prompts.txt

Looking for older digests? Browse all digests here.