Daily Digest

Daily Digest – 10 June 2026

2026 Capabilities Progress

  1. Page generation — Canvas merged full-screen extension support [24][112][136] and advanced work on entity reference selection APIs [130][166], both foundational for AI-driven page assembly. Component context improvements [132] and orchestration loop work [134][174] are in progress but not yet production-ready.
  2. Context management — The Context Control Center completed a redesigned listing interface [3][184] and simplified the underlying data model [177][185], making context organization more accessible. Documentation gaps around tagging [182] and architectural discussions on multi-type context support [179][180] indicate the feature set is still evolving.
  3. Background agents — Significant architectural work is underway with the agent entity structure migration into AI core [7][8][9], including runner abstractions [49][50][56], event systems [51], permissions controls [57], and message queue integration [59]. This represents planning and scaffolding; no background execution capabilities were delivered today.
  4. Design system integration — No direct progress today. The meta issue [24] remains open with no recent activity, and no component generation capabilities were advanced during this period.
  5. Content creation and discovery — AI Search added filterable search tools with attribute-level precision [193] and non-content entity access controls [192]. Content editing improvements include CKEditor streaming fixes [5][26], image description support for automators [79], and guardrail enhancements [33][41]. Translation quality improved with prompt warning fixes [198][207] and field extraction corrections [17][18][199][209].
  6. Advanced governance — Canvas delivered conflict detection for auto-save operations [122][165][167] and selective publishing with resolution UI [175], plus translation audit trail improvements [116][160]. Prompt validation [14] and automator event hooks [38] remain in development.
  7. Intelligent website improvements — No direct progress today. Observability work to align with GenAI telemetry standards [27][39] and evaluation framework development [31] represent instrumentation groundwork but no learning or proposal capabilities were delivered.
  8. Multi-channel campaigns — No direct progress today. No work on cross-channel content distribution, social media integration, or email platform connectivity was observed during this period.

Shipped

  1. AI module maintenance releases published — Versions 1.4.2 and 1.3.7 were released after successful quality assurance testing [12][62].
  2. AI automator form crash fixed — Resolved a bug where adding automator fields to entity bundles with no available fields caused the configuration form to break [15][70][71].
  3. Canvas translation management completed — Translation features for configuration-based component trees were merged, enabling TMGMT integration for content templates and page regions [121].
  4. Canvas automated dependency updates enabled — Renovate integration was merged to automatically update npm and composer dependencies, reducing maintenance overhead [113][169][172].
  5. AI translate installation warning eliminated — Fixed undefined array key errors appearing on fresh installations when no language-specific prompts were configured [198][207].

Ongoing

  1. AI agent architecture migration to core module — Work continues on moving agent definitions from the contrib ai_agents module into core AI, introducing pluggable runner systems supporting both Drupal-native and Symfony AI execution [7][8][9].
  2. Canvas upgrade bug blocking 1.5.1 adoption — A critical issue causing LogicException errors on content templates with dynamic field bindings during upgrades from 1.4.1 to 1.5.1 is under review [111][142].
  3. AI ecosystem documentation catalog in progress — A curated index listing all AI-related modules with links and summaries is being developed to improve discoverability for new adopters [25][85].
  4. Anthropic provider gaining prompt caching and PDF support — A merge request adding cost-saving prompt caching and PDF document processing capabilities is under review [91].
  5. Canvas full-screen extension type in development — Work has begun on supporting standalone tools like translation dashboards, addressing limitations of the current sidebar-only model [112][136].

AI (Artificial Intelligence)

The AI module saw significant quality and stability improvements during this period, with two releases shipped and several structural bugs resolved. The team also advanced planning for version 2.0, particularly around agent architecture, while clarifying the ecosystem for adopters.

Two maintenance releases—1.4.2 and 1.3.7—were published after successful QA runs. Contributors fixed a critical testing issue where the MDX editor's conversion to a Drupal behavior caused JavaScript test failures (Fix FunctionalJavascript test failures), merged by Artem Dmitriiev. A longstanding form crash when automator fields were added to entity bundles with no fields was also resolved (Add AI Automator field form breaks), with merge requests by shalini jha and Artem Dmitriiev merged to multiple branches.

Work continued on the [META] Move AI Agents architecture into the AI module, led by Marcus Johansson with Michael Lander and harivansh sharma assigned. This restructuring will move agent definitions from the contrib ai_agents module into core AI, introduce a pluggable runner system, and support both Drupal-native and Symfony AI execution backends. Related planning issues cover structured input/output, control mechanisms, and migration paths. A discussion issue explores alignment with other agent frameworks to inform the canonical agent definition (Define the canonical agent definition).

Bug reports and feature requests arrived across chatbot, automator, and provider components. Two open merge requests address token usage loss in streamed OpenAI chat output (Fiber branch drops token usage, by Jérôme Tchania) and improved media rendering in the chatbot viewport (Render images and videos inside the chatbot viewport, by Ahmad Khalil). A Gemini streaming bug with 2.x models (streamGenerateContent causes 400 Bad Request) remains open.

Marcus Johansson filed a documentation task to create a curated ecosystem index listing all related modules—providers, tools, agents, and distributions—to improve discoverability for new adopters (Document the AI module ecosystem), assigned to Rob Loach. The team also began work to align observability tracing with OpenTelemetry GenAI conventions (Align ai_observability to gen_ai.* conventions) and structured a cross-project Drupal Eval Commons framework to enable shared evaluation formats across AI modules (Drupal Eval Commons umbrella).

How can I help on this project?

  • Prioritize design-system alignment work for Component Generation (assigned to Rakhi Mandhania, [META]Design System Integration) to reduce structural risk as code-generation features mature.
  • Allocate resources to finalize the agent architecture migration plan and shepherd the related issues through review, as this is foundational to the 2.0 release.
  • Support ecosystem documentation efforts by ensuring the curated module index is published promptly, as this directly lowers the barrier to adoption for new users.

View project on GitLab: ai

Drupal AI Initiative

The AI Initiative made significant progress consolidating and finalising planned features this week. Five proposed capabilities moved to closed status as the team wrapped up planning work: content pre-moderation using AI, document classification, AI-powered content suggestions for editors, image safety validation to block inappropriate uploads, and a marketing article template. These closures indicate the team has completed definition work and is ready to move forward with delivery.

Two new ideas emerged. Marcus Johansson (Marcus_Johansson) proposed an AI Cost & Usage module that would give site owners visibility into their AI spending through a local dashboard, with optional community telemetry to help the initiative understand real-world usage patterns. A second proposal suggests creating a provider for Vercel AI Gateway, which would let sites access hundreds of AI models through a single endpoint rather than managing multiple vendor relationships.

Development continues on audio transcription and summarization, assigned to Artem Dmitriiev (a.dmitriiev), which will automatically enrich audio media with text transcripts and summaries.

How can I help on this project?

  • Prioritise whether cost visibility and community telemetry warrant immediate resourcing, as this directly addresses site owner concerns about AI spending transparency.
  • Evaluate the strategic value of gateway provider integration versus direct provider connections to inform roadmap decisions.

View project on GitLab: ai_initiative

AI Integration - ECA

Work is underway to extend the module's integration with artificial intelligence capabilities. Marcus Johansson (Marcus_Johansson) has begun implementing support for Add ECA execute action for Rerank, which will enable users to improve search result relevance through AI-powered reordering of content. This capability allows the system to automatically prioritize the most relevant information when users perform searches, enhancing the user experience without requiring manual sorting or filtering.

The work is part of the project's roadmap toward a stable 1.0.0 release and aligns with broader AI innovation initiatives. Currently, the module supports several AI operations including conversational interfaces, content moderation, and speech processing. Adding reranking functionality closes a gap in the available capabilities and brings the module closer to feature completeness.

The issue was opened within the reporting period and is assigned for implementation. No code has been merged yet, which is expected for work that has just begun.

How can I help on this project?

  • Prioritize resources or clear roadblocks if the 1.0.0 release timeline is critical to organizational AI strategy.
  • Identify early adopters within the organization who could test reranking features once available and provide business-relevant feedback.

View project on GitLab: ai_integration_eca

AI Agents Test

The AI Agents Test module saw automated compatibility work this period in preparation for Drupal 12. A merge request was opened to address deprecated code, part of the broader Automated Drupal 12 compatibility fixes effort. This work ensures the module will continue functioning when the Drupal platform upgrades to its next major version, avoiding disruption for users who depend on this testing functionality.

No code was finalized or merged during this period. The compatibility updates remain under review as an open merge request. This is routine maintenance work typical for modules in active development, ensuring the project stays aligned with the broader Drupal ecosystem's evolution.

How can I help on this project?

  • Assign a technical reviewer to evaluate and merge the pending Drupal 12 compatibility updates, ensuring the module remains supported.
  • Clarify the project's roadmap and priority level to help contributors understand whether to invest in further development or focus on maintenance.

View project on GitLab: ai_agents_test

Anthropic Provider

Development is underway to add two significant capabilities to the Anthropic integration. A new merge request for Bump SDK constraint and drop method_exists defensive check was opened by Carlos Ospina (camoa), bringing prompt caching and PDF document support to the module.

Prompt caching will allow the system to reuse conversation context efficiently, reducing costs and improving response times for repeated interactions. Site administrators will be able to enable this feature through a simple toggle and choose how long to cache prompts (5 minutes or 1 hour). The work also introduces the ability to process PDF documents directly, expanding the types of content the system can work with beyond plain text.

This represents a second phase of modernization work building on earlier improvements. The changes are designed for the module's 1.3 release branch and are ready for review.

How can I help on this project?

  • Ensure the maintainer team has capacity to review and merge the pending work, which spans over 1,600 lines of changes.
  • Consider whether beta-testing resources are available to validate prompt caching before the 1.3 stable release.

View project on GitLab: ai_provider_anthropic

Huggingface Provider

The Huggingface Provider module saw development progress during this period with work beginning on text translation capabilities. Ahmad Khader (Ahmad-Khader) opened a merge request to add text translation, representing 132 lines of changes. This enhancement would allow users to translate text content through the Huggingface platform, expanding the module's utility for multilingual websites and global content operations.

The merge request is currently under review and has not yet been integrated into the module. No other development activity, commits, or issue updates were recorded during this reporting period.

How can I help on this project?

  • Ensure the translation feature receives timely code review so it can be tested and released to users who need multilingual capabilities.
  • Consider whether additional language or AI provider features should be prioritized to maximize the module's value for content operations.

View project on GitLab: ai_provider_huggingface

AI Agents

The AI Agents project saw active issue triage and community contribution over the past 24 hours, with focus on quality improvements and alignment with the broader AI module ecosystem.

The team closed a duplicate bug report and addressed a labeling inconsistency in the administration menu, though the fix to align with the AI module's new "Tools & Automation" categorization is still in progress. Several data handling bugs were identified and accepted: issues with status reporting when tool output is missing expected metadata, entity field information incorrectly inheriting properties across fields, and the content seeding tool mishandling complex field structures. Contributors have proposed solutions for most of these.

Two issues address important guardrails: ensuring complete tracking information for tool execution and preventing configuration conflicts when custom agents use names that clash with built-in functionality. The latter is marked ready for final review.

Work continues on improving the configuration interface for property constraints and on compatibility with the upcoming Drupal 12 release.

How can I help on this project?

  • Prioritize review capacity for the configuration collision fix, which is ready for final approval and prevents admin errors.
  • Ensure testing resources are available for the data handling bugs, which affect content creation workflows and multi-field entities.

View project on GitLab: ai_agents

Drupal Canvas

Canvas continues progress on multilingual capabilities and core platform stability. The most significant delivery is the completion of translation management for configuration-based component trees: Symmetrically translatable config-defined component trees, STEP 4 was merged, enabling TMGMT integration for content templates and page regions. This complements the content-entity translation work completed last week and brings Canvas to parity across both content and configuration translation workflows.

The team merged automated dependency updates via Renovate, reducing ongoing maintenance overhead. Several translation UI refinements are in review, including a delete translation confirmation modal and fixes for missing language checkmarks in templates.

A critical bug affecting users upgrading from 1.4.1 to 1.5.1 is under review: LogicException on content templates with dynamic field bindings. A merge request has been opened and requires validation before release. Additionally, work has begun on a full-screen extension type to support standalone tools like translation dashboards, addressing feedback that the current sidebar-only model constrains certain workflows.

Open testing work includes multilingual search indexing and improved Canvas AI orchestration to avoid timeouts on hosted platforms.

How can I help on this project?

View project on GitLab: canvas

Context Control Center (CCC)

Design and architecture discussions dominated the past 24 hours as the team weighs strategic decisions about the module's structure and usability.

Two significant pieces of work are now ready for review. Matt Glaman (mglaman) has prepared an updated listing page design that reorganizes how context information is displayed to users, making status and scope information more visible and accessible. He has also delivered work to simplify the underlying data structure by removing unnecessary complexity around how context items are organized, following improvements in the system's scheduling capabilities.

Several conversations are underway about the product's direction. The team is exploring whether to support multiple types of context items to enable features like providing good and bad content examples. Michael Anello (ultimike) raised questions about how the tagging feature works and requested clearer documentation. Separately, Kristen Pol (kepol) is investigating whether less expensive language models can deliver acceptable results for the Chicago demonstration, potentially reducing costs from $1-2 per page to more sustainable levels.

How can I help on this project?

  • Prioritize a decision on whether multiple context item types are needed, as this affects work already in progress on simplification.
  • Allocate budget for improved user documentation, particularly around the tagging feature where gaps have been identified.

View project on GitLab: ai_context

AI Dashboard

The AI Dashboard project made significant progress over the past 24 hours on improving how users discover and adopt AI capabilities within Drupal. Two major updates were delivered focusing on making the AI ecosystem more accessible.

The team completed work on AI Ecosystem Documentation, creating a browsable catalog of all AI-related modules available in the Drupal ecosystem. This new resource shows module titles, descriptions, and adoption metrics (install counts), helping organizations evaluate which AI tools best fit their needs. The catalog has been organized by categories to make navigation easier.

New content moderation and image safety validation capabilities were also added to the platform, expanding the range of AI-powered features available to users.

One issue remains open: users report that the "View Commands" buttons in the Features section are missing required installation information, preventing them from deploying AI recipes to their sites. Additionally, the team has identified several refinements needed for the ecosystem catalog, including improved readability and automated updates from the central module registry.

How can I help on this project?

  • Prioritize fixing the installation command issue, as it currently blocks users from deploying any AI features to their sites.
  • Consider allocating resources to automate the ecosystem catalog updates, ensuring users always see current module information without manual maintenance overhead.

View project on GitLab: ai_dashboard

Field Widget Actions

Field Widget Actions is a Drupal module in active development. Over the last 24 hours, the team made progress on quality improvements related to how users interact with the system's action selection interface.

A bug was identified and assigned where certain actions appear as options in selection menus even when they cannot actually be used in that context. This creates confusion when users attempt to add functionality that the system will ultimately reject. Tamás Brückner (brtamas) has been assigned to resolve this issue, which will improve the user experience by ensuring that only genuinely available options are presented in dropdown lists. This work is part of the AI Initiative Sprint, indicating alignment with the project's strategic direction in automation capabilities.

No code changes were completed during this period, which is normal for a module in active development where issue diagnosis and assignment precede implementation work.

How can I help on this project?

  • Ensure the assigned developer has adequate time allocated to complete the interface bug fix without competing priorities.
  • Clarify whether the AI Initiative Sprint has specific delivery milestones that should inform prioritization of this and related work.

View project on GitLab: field_widget_actions

AI translate

The AI translate module made progress on stability improvements and editor experience over the past 24 hours. A fix was merged that eliminates warning messages that appeared on fresh installations when no language-specific translation prompts had been configured yet. This issue, Undefined array key warning in ChatTranslationProvider when no language-specific prompt is configured, was contributed by Matthew Tift (mtift) and removes technical errors that could confuse users during initial setup.

Several improvements remain in progress. Work continues on making the translation interface more user-friendly, including a proposal to display human-readable model names instead of technical identifiers when editors select AI translation options. Contributors are also addressing issues where certain content types are not being translated correctly, including URL aliases and custom fields that don't follow standard naming conventions.

The module is also preparing for future Drupal versions, with automated compatibility updates being generated for Drupal 12. As an evolving module, it has several open issues around field support and error handling that are normal for this stage of development and are being actively worked on by contributors including Jordan Koplowicz (koppie), yuri.p (rurik666), and others.

How can I help on this project?

  • Consider funding dedicated testing resources to validate translation accuracy across diverse content types and field configurations, particularly for common modules like Address and Metatag.
  • Facilitate coordination between this project and the AI core module team to align on shared infrastructure decisions and prioritize feature requests.

View project on GitLab: ai_translate

Pinecone VDB Provider

During this reporting period, automated compatibility work began to prepare the module for Drupal 12. The Automated Drupal 12 compatibility fixes for ai_vdb_provider_pinecone 2.0.x-dev issue was opened, with a corresponding Automated Project Update Bot fixes merge request submitted. This work ensures the module will remain compatible with the next major version of Drupal, preventing future disruption when the platform upgrades.

This is routine maintenance expected for modules at this stage. The automated changes remove outdated code patterns and align the module with current platform standards. No code has been merged yet, so this represents work in progress rather than delivered functionality. There are no blocking issues at this time.

How can I help on this project?

  • Allocate developer time to review and merge the pending compatibility changes, ensuring the module stays current with platform evolution.
  • Consider prioritizing manual testing resources once automated fixes are merged to validate functionality on Drupal 12.

View project on GitLab: ai_vdb_provider_pinecone

SQLite VDB Provider

Activity over the past 24 hours has been limited to automated maintenance work. A merge request titled Automated Project Update Bot fixes was opened by Project Update Bot (Project-Update-Bot) to apply routine updates to the project. This type of automated maintenance helps keep the module's dependencies and coding standards current, reducing technical debt and ensuring compatibility with the broader Drupal ecosystem.

No issues were updated, and no code was merged during this period. This reflects a quiet day in the project's development cycle, with no user-facing changes or feature progress to report.

How can I help on this project?

  • Review the pending automated update merge request to ensure it can be merged promptly, keeping the project current with ecosystem standards.
  • Evaluate whether the project needs additional development resources to accelerate feature delivery or address user needs.

View project on GitLab: ai_vdb_provider_sqlite

AI usage limits

The AI usage limits module saw minimal activity over the past 24 hours. An automated maintenance task was initiated to apply routine fixes and updates to the project infrastructure, opened by the Project Update Bot. This type of housekeeping work helps keep the module aligned with current platform standards and ensures ongoing compatibility.

No substantive feature development, bug fixes, or user-facing improvements were delivered during this period. The module remains in active maintenance, with no urgent issues or blocking concerns identified.

How can I help on this project?

  • Consider evaluating whether additional developer resources are needed to accelerate feature development or address any backlog of user requests.
  • Engage with stakeholders to clarify strategic priorities for the module and ensure the team has clear direction on what capabilities matter most.

View project on GitLab: ai_usage_limits

References

[3] Update documentation for FWA and ai_automators (ai, Issue)

[5] ai_ckeditor: streamGenerateContent causes 400 Bad Request with Gemini 2.x models (ai, Issue)

[7] Move the agent entity structure into the AI module (ai, Issue)

[8] [META] Move AI Agents architecture into the AI module (ai, Issue)

[9] [Discuss] Define the canonical agent definition (informed by other agent frameworks) (ai, Issue)

[12] Release 1.4.2 and 1.3.7 (ai, Issue)

[14] ai_prompt entity validation (ai, Issue)

[15] Add AI Automator field form breaks if fields not available in any bundle (ai, Issue)

[17] Remove hardcoded field-name patterns from ai_translate's TextExtractor::shouldExtract() (ai, Issue)

[18] Move TextExtractor service to AI core (ai, Issue)

[24] [META]Design System Integration with Component Generation (ai, Issue)

[25] Document the AI module ecosystem: a categorized list of related modules with links and summaries (ai, Issue)

[26] CKEditor AI streaming fails for OpenAI when Guzzle uses StreamHandler for "stream => true" requests (ai, Issue)

[27] [Layer 2] Cross-project result envelope tracking (every_eval_ever 0.2.2 + additional_details) (ai, Issue)

[31] Drupal Eval Commons umbrella (ai, Issue)

[33] Rename Guardrails "RegExp guardrail" plugin to just "RegEx" (ai, Issue)

[38] Ai Automators could need more Events before asking AI (ai, Issue)

[39] [Layer 2] Align ai_observability OTel emission to GenAI semantic conventions (gen_ai.*) (ai, Issue)

[41] Add a 'During generate' (streaming) section to the Guardrail Set form and let guardrails declare their supported modes (ai, Issue)

[49] Create an agent runner abstraction as plugins (ai, Issue)

[50] Create a Drupal agent runner (ai, Issue)

[51] Shared event points for both agent runners (ai, Issue)

[56] Create a Symfony AI agent runner (ai, Issue)

[57] Agent runners must support running tools under a configured role/permissions (ai, Issue)

[59] Integrate the runner with Symfony Messenger (or similar) (ai, Issue)

[62] QA 1.4.1 and 1.3.7 (ai, Issue)

[70] Issue #3558728: Fix AI Automator form break when bundle has no fields (ai, MR)

[71] Issue #3558728: Fix AI Automator form break when bundle has no fields (ai, MR)

[79] Issue #3586478: "Add rich-text image description support to AI Automators" (ai, MR)

[85] Create article template for additional current capabilities (ai_initiative, Issue)

[91] Issue #3590963: Bump SDK constraint and drop method_exists defensive check (ai_provider_anthropic, MR)

[111] LogicException on content templates with dynamic field bindings and config schema validation errors on fallback components after upgrading from 1.4.1 to 1.5.1 (canvas, Issue)

[112] Add "page" extension type for full-screen Canvas extensions (canvas, Issue)

[113] CI: Renovate job to auto-update npm and composer dependencies (canvas, Issue)

[116] Update Canvas component auditing UI to include translation labels (canvas, Issue)

[121] Symmetrically translatable config-defined component trees, STEP 4: allow translating component instance inputs via TMGMT for ContentTemplates & PageRegions (canvas, Issue)

[122] Detect conflicts for Page entities during the handling of auto-saves/pending calls (canvas, Issue)

[130] Backend APIs for Entity Reference Selection (canvas, Issue)

[132] Canvas AI: Component context required props missing required flag (canvas, Issue)

[134] Canvas AI: Client-side orchestration loop for Canvas AI(POC) (canvas, Issue)

[136] Add "page" extension type for full-screen Canvas extensions (canvas, MR)

[142] bug: #3591642 Fix LogicException on component instance updates and config... (canvas, MR)

[160] #3591612 auditing UI to include translation labels (canvas, MR)

[165] fix(Conflict detection): Exclude computed internal fields from Page entity normalization (canvas, MR)

[166] feat: #3585354 add content-entity-types / fields endpoints for entity reference selection UI (canvas, MR)

[167] feat(Conflict detection): #3589076 "Detect conflicts during auto-save polling" (canvas, MR)

[169] Resolve "CI: Renovate job to auto-update npm and composer dependencies" (canvas, MR)

[172] Issue #3590179: Add Renovate to auto update dependencies (canvas, MR)

[174] POC-Phase -1 (canvas, MR)

[175] Add conflict detection and resolution UI for selective publishing (canvas, MR)

[177] Remove AiContextItemType (ai_context, Issue)

[179] [Discuss] Usage of "Context scope" and "Target bundles" (ai_context, Issue)

[180] [Discuss] Context item content entity bundles (ai_context, Issue)

[182] Context scope: Tags (documentation) (ai_context, Issue)

[184] Issue #3586219: Update context listing page per new design (ai_context, MR)

[185] Issue #3584775: Remove AiContextItemType and make ai_context_item bundleless (ai_context, MR)

[192] Issue #3579758: Provide API for access checks on non-content entities (ai_search, MR)

[193] Issue #3584010: Add FilterableRagTool with per-index derivatives and attribute filtering (ai_search, MR)

[198] Undefined array key warning in ChatTranslationProvider when no language-specific prompt is configured (ai_translate, Issue)

[199] Custom fields without field_ prefix are not extracted for translation (ai_translate, Issue)

[207] Issue #3585520: Undefined array key warning in ChatTranslationProvider when no language-specific prompt is configured (ai_translate, MR)

[209] Check field definition for translation, instead of field name (ai_translate, MR)

Period: 10 June 2026 to 11 June 2026  |  Generated: 11 June 2026 05:00 GMT

To view the full AI prompts used to generate this digest check this file: 24h_2026-06-10_prompts.txt

Looking for older digests? Browse all digests here.