Daily Digest

Daily Digest – 7 July 2026

2026 Capabilities Progress

  1. Page generation — Validation improvements ensure Canvas AI now rejects invalid component configurations, improving reliability of AI-generated pages [113]. Client-side orchestration work continues to address timeout issues on hosted platforms [71][72].
  2. Context management — The Context Control Center merged accessibility improvements and documentation fixes [143][144], with accessibility improvements in progress [141] and performance optimizations to prevent unnecessary catalog scans under review [145]. Integration of context into CKEditor AI tools progressed [148], with beta 3 release planning and quality improvements in progress.
  3. Background agents — New agent skills are under review to enable programmatic configuration of AI settings and field automators [5][8][16][17]. The AI-ECA interceptor submodule draft enables workflows to monitor and intervene in AI interactions [35][36], supporting automated governance scenarios.
  4. Design system integration — Canvas merged fixes ensuring component libraries can evolve without breaking existing pages, particularly on multilingual sites where component property changes previously caused validation failures [103]. Heuristics were added to streamline which content fields appear during component configuration [96].
  5. Content creation and discovery — A new AI reranking processor for Search API is under review to improve semantic search relevance [4][15]. Batch embedding indexing work continues [3], and Experience Builder added search capabilities for its content [33]. CKEditor AI tools received fixes for tone adjustment and summarization features [7][9][23-26][28][30].
  6. Advanced governance — Canvas completed a rudimentary conflict resolution UI allowing editors to visually compare and resolve conflicts between drafts and published content [63][105]. Security hardening for execution principals in agent API calls was merged [13][32], and usage record access controls were strengthened [133][155].
  7. Intelligent website improvements — No direct progress today.
  8. Multi-channel campaigns — The AI translate module refactored its architecture to support agent-driven translation through a unified orchestrator service, enabling automated translation workflows beyond manual triggers [157][158][159].

Shipped

  1. AI Module CKEditor integration fixes — Fixed hardcoded error messages so each CKEditor tool displays appropriate guidance when no text is selected, and resolved fatal errors affecting Drupal 11.4 compatibility [7][9][23][24][25][26][28][30].
  2. AI Agents tool restriction enforcement — Fixed a bug where forced tool values were ignored when properties remained visible, ensuring agents now behave as configured [46][49].
  3. Canvas translation infrastructure improvement — Resolved issue where deleting optional component properties broke existing multilingual pages by tracking translation requirements per component version [61][103].
  4. Canvas conflict resolution interface — Delivered visual tool enabling editors to compare conflicting versions side-by-side and choose which to keep [63][105][110].
  5. Context Control Center accessibility and quality improvements — Merged fixes for form and table accessibility, resolved test failures from external dependency changes, and corrected documentation inconsistencies [122][123][125][141][143][144][156].

Ongoing

  1. AI Module Search API reranking processor — Development underway to enable AI-powered reranking of search results for improved relevance across all Search API backends [4][15].
  2. AI Module agent configuration tools — Command-line tools under review to enable scripted configuration of AI settings and field automators for deployment automation [5][8][16][17].
  3. Canvas headless preview authentication — Work in progress to establish security model for headless draft preview, currently blocked pending security review [64][102].
  4. OpenAI Provider migration to Responses API — Transition underway to newer OpenAI interface to maintain compatibility as older Chat Completions API is phased out [38].
  5. AI translate refactoring for agent integration — Core logic being restructured into shared service layer to support agent-driven automated translation workflows [157][158][159].

AI (Artificial Intelligence)

The AI module delivered several user-facing fixes to its rich text editor (CKEditor) integration. A longstanding issue where all AI CKEditor tools displayed the same "you must select text to summarize" message has been resolved; each tool (Translate, Tone, SpellFix) now shows appropriate context-specific guidance when no text is selected. A related fix ensures that plugin configuration fields are hidden entirely until the user selects text, reducing confusion. Additionally, a compatibility workaround was applied to address fatal errors affecting CKEditor tools on Drupal 11.4.0.

Work is progressing on new automation capabilities. Merge requests are under review to add command-line tools (called "agent skills") that will allow site operators to configure AI settings, guardrails, and field automators programmatically rather than only through the administrative interface. These tools are intended to support scripted deployment and integration with AI-driven site management workflows.

A new Search API processor for reranking search results using AI models is in development. Reranking improves the relevance of semantic search by re-ordering results after an initial query; the processor will work with any Search API backend (database, Solr, or vector stores). Separately, work continues on embedding performance improvements and code quality tasks, including resolution of static analysis issues in the API Explorer component.

How can I help on this project?

  • Review and approve the Search API reranking processor merge request to enable the feature for semantic search users.
  • Prioritize a decision on how the module should notify site administrators when AI providers release better models, as this affects long-term maintainability and user experience.

Contributors: Abhisek Mazumdar (abhisekmazumdar), Ann Mary Sruthy (annmarysruthy), Artem Dmitriiev (a.dmitriiev) [2], Daniel Rodriguez (danrod), scott falconer (scottfalconer)

View project on GitLab: ai

Experience Builder

A new merge request was opened to create a configurable search api processor for XB data. This work will enable better search functionality for content created with Experience Builder, making it easier for site visitors to find relevant information. The change is currently under review and represents progress toward improving the discoverability of content built using the module.

No other activity was recorded during this period. The module continues its development toward production readiness.

How can I help on this project?

  • Ensure the team has access to review and testing resources to move the search processor work toward completion.
  • Clarify whether improved search functionality aligns with near-term product priorities and communicate any timeline expectations to the team.

Contributors: Luhur Abdi Rizal (el7cosmos)

View project on GitLab: experience_builder

AI Integration - ECA

Work this period focused on extending the module's coverage of AI operations and introducing capability for custom workflows to monitor and intervene in AI provider interactions.

A new work item was opened to add support for the Summarization operation, which will allow automated workflows to trigger text summarization tasks. This fills a gap in the module's current operation coverage and has been assigned for development.

A draft merge request was opened for the AI ECA Interceptor submodule, which will expose AI request and response events to workflow automation. This feature will enable organizations to inspect, log, transform, or block AI interactions before they reach providers or after responses are received. The capability is particularly relevant for governance, compliance monitoring, and custom error handling around rate limits and content safety violations.

Both initiatives are tagged as part of the road to version 1.0.0, indicating progress toward production readiness.

How can I help on this project?

  • Clarify business priorities between operation coverage and interceptor capabilities to help guide which features reach 1.0.0 first.
  • Ensure development resources remain allocated through the 1.0.0 milestone to sustain current momentum.

Contributors: Marcus Johansson (Marcus_Johansson)

View project on GitLab: ai_integration_eca

OpenAI Provider

The OpenAI Provider module made progress this period on future-proofing its integration with OpenAI's platform. A significant update was merged to add support for file handling capabilities, broadening the types of content the module can work with. This enhances the module's utility for workflows that involve documents or images alongside text-based interactions.

A second update is in progress to migrate the module's chat functionality to OpenAI's newer Responses API. OpenAI is phasing out its older Chat Completions interface in favor of Responses, which will eventually become the standard. This work ensures the module continues to function reliably as OpenAI evolves its platform, avoiding disruption for users who depend on chat features. The approach aligns the module with how other providers are implemented, maintaining consistency across the broader integration framework.

Both efforts represent strategic investments in the module's long-term viability and feature set, positioning it to support emerging use cases and platform changes.

How can I help on this project?

  • Prioritize testing and review resources for the Responses API migration to accelerate its completion before OpenAI deprecates the older interface.
  • Assess whether documentation or training materials need updating to reflect new file handling and API capabilities for end users.

Contributors: Ahmad Khalil (ahmad-khalil-imagex), Bryan Sharpe (b_sharpe)

View project on GitLab: ai_provider_openai

Google Vertex Provider

A significant merge request was opened to refactor the connection layer and add test coverage to the module. This work aims to improve the reliability and maintainability of the Google Vertex integration, making it easier to identify and prevent issues before they affect users. Testing infrastructure is a foundation for stable releases and reduces the risk of regressions as the module evolves.

The Refactor HTTP client & add tests merge request represents substantial progress toward production readiness, though the work is still under review and has not yet been integrated.

No issues were updated and no code was merged during this period. Activity remains focused on preparing the testing and architectural improvements for integration.

How can I help on this project?

  • Prioritize review capacity to evaluate and merge the open testing work, which is essential for stable release readiness.
  • Consider resourcing to support ongoing quality assurance efforts as the module matures toward production use.

Contributors: Tamás Nagy (yce)

View project on GitLab: ai_provider_google_vertex

Milvus VDB Provider

The Milvus VDB Provider module saw progress on reliability improvements during this period. Work addressing connection stability issues was merged, specifically targeting scenarios where the Milvus service becomes unresponsive while the rest of the hosting environment remains functional. This resolves interruptions that could prevent users from accessing vector search capabilities.

Additionally, a new proposal has been opened to address broader connection failure handling. This work aims to make the integration more resilient when communication with the Milvus database is disrupted, reducing downtime and improving the user experience for sites relying on vector-based search and retrieval.

These improvements strengthen the module's production readiness by addressing operational stability, which is essential for maintaining consistent service quality as adoption grows.

How can I help on this project?

  • Prioritize testing resources to validate the connection failure fixes in real-world deployment scenarios, especially diverse hosting environments.
  • Ensure the team has capacity to review and finalize the open connection handling work to maintain momentum on reliability improvements.

Contributors: Remco Hoeneveld (RemcoHoeneveld), Scott Euser (scotteuser) [1]

View project on GitLab: ai_vdb_provider_milvus

Model Context Protocol

A new merge request has been opened to add support for write operations, specifically create and update functionality. This capability would allow users to not only read data through the Model Context Protocol module, but also to create new content and modify existing content. This represents a significant expansion of the module's capabilities, moving it from a read-only integration to a bi-directional system that can actively manage content within Drupal.

The merge request is currently under review. No issues were updated and no code was committed to the main codebase during this period.

How can I help on this project?

  • Ensure the team has dedicated code review capacity to evaluate the write operations merge request, as this feature significantly expands the module's scope and risk profile.
  • Clarify the security and permission requirements for write operations to guide development decisions.

Contributors: Shubham Goel (shubhamgoel)

View project on GitLab: mcp

AI Agents

The project resolved a bug affecting how agents enforce tool restrictions. Previously, when administrators configured a tool to use a forced value but left the property visible, the restriction was silently ignored and the AI model could call the tool with any value it chose. This has now been fixed, ensuring that forced values are respected regardless of visibility settings. The correction means agents will now behave as configured, improving reliability and control over tool behavior.

Work is underway to improve the user experience when configuring agents. A change is being reviewed that would replace the plain text box for structured output schemas with a dedicated code editor featuring syntax highlighting and validation, making it easier to define and maintain complex output formats without errors.

Two additional issues were opened: one addressing a crash in the debugging interface when certain tool responses lack expected metadata, and another questioning whether the artifact storage checkbox is necessary given observed behavior. Both have proposed solutions under review.

How can I help on this project?

  • Prioritize review and testing of the code editor improvement, which will reduce configuration errors and support faster agent development.
  • Clarify the intended behavior of artifact storage to resolve user confusion and ensure documentation aligns with actual functionality.

Contributors: Ahmad Khader (Ahmad-Khader), AKHIL BABU (AkhilBabu), Prabhavathi Vanipenta (prabha1997)

View project on GitLab: ai_agents

Tool API

The Tool API project saw limited activity during this period. An issue titled File Input Handling was opened to investigate how the module should handle file inputs when tools are invoked by AI agents. This research task focuses on determining secure approaches for accepting and processing files, including input formats, file loading, and normalization. The work is exploratory in nature and will inform future design decisions about whether file handling capabilities should be integrated into this module or developed separately.

No code changes were merged or committed during the reporting period. As an early-stage module supporting AI integration, ongoing research and planning work like this is typical and helps ensure the foundation is built correctly before broader feature development begins.

How can I help on this project?

  • Clarify strategic priority: Confirm whether file handling for AI agents is a near-term requirement or can be deferred based on adoption patterns.
  • Ensure research findings are reviewed: Allocate time for decision-makers to evaluate outcomes once the proof of concept is complete.

View project on GitLab: tool

Drupal Canvas

Over the past 24 hours, Drupal Canvas made significant progress on core infrastructure for translation, conflict resolution, and AI-powered page building, with three major integrations completed.

A fix for translation support was merged, addressing a critical issue where deleting optional component properties broke existing pages on multilingual sites. The system now tracks each property's translation requirements per component version, ensuring smooth evolution of component libraries without disrupting live content.

The rudimentary conflict resolution interface was completed, giving editors a visual tool to resolve conflicts between their draft changes and published content. Users can now compare conflicting versions side-by-side and choose which to keep, reducing the risk of lost work when multiple editors touch the same page.

Work also merged to refine which content fields appear in component configuration, filtering out internal metadata and duplicate options to streamline the editing experience. Additionally, a validation fix for Canvas AI prevents the system from accepting invalid component configurations, improving reliability of AI-generated pages.

Several open issues continue to evolve, including authentication infrastructure for headless preview, broader conflict resolution capabilities, and client-side AI orchestration to avoid timeouts on hosted platforms.

How can I help on this project?

Contributors: **** (project_19391_bot_cb443e60680723bba08a5351bf595251), Abhishek Dhariwal (adaptative) [1], Ben Mullins (bnjmnm), Bálint Kléri (balintbrews), Christian López Espínola (penyaskito) [2], Feliksas Mazeikis (f.mazeikis) [1], Fons Vandamme (f0ns), Jakob P (japerry), Lauri Timmanee (lauriii), project update bot (Project-Update-Bot), Ravi Maniyar (ravi.maniyar.123) [1], scott falconer (scottfalconer), Stephanie Galata (sgalata), Wim Leers (wimleers)

View project on GitLab: canvas

Context Control Center (CCC)

The Context Control Center module made progress on quality improvements and release preparation during the reporting period. Two merge requests were completed: one addressing form and table accessibility, including replacing disabled read-only select elements with plain text and improving screen reader feedback, and another fixing documentation inconsistencies around default configuration values and outdated API references.

A test failure was also resolved. The module's functional tests had begun failing due to a change in an external dependency (the AI Agents module), which altered an administrative URL without notice. The hardcoded path in the test was updated to restore CI pipeline stability.

Work continues on the beta 3 release. A gap analysis comparing the current beta API against requirements for a stable 1.0 release was completed and closed, generating several follow-up issues. Multiple merge requests remain in progress addressing cacheability, access control, performance (particularly avoiding full-catalog scans for agents with no scope subscriptions), and further accessibility improvements. Several issues are exploring integration opportunities with related modules in the ecosystem, such as the Annotations and DUADP modules, to ensure the module supports emerging use cases.

The project remains in active development toward beta 3, with planning underway for release candidate and stable versions.

How can I help on this project?

  • Prioritize review and merge decisions on the eight open merge requests currently awaiting maintainer feedback to unblock beta 3 release preparation.
  • Clarify strategic direction on integrations with the Annotations and DUADP modules to guide whether follow-up work should be scoped into the stable 1.0 roadmap or deferred.

Contributors: Jesse Hofmann-Smith (jessehs), Joshua Fernandes (joshua1234511), Juan Correa (jucs7) [1], Kristen Pol (kepol) [1], Pravesh Poonia (Pravesh_Poonia) [1], Rob Loach (robloach), Tamas Balog (tbalog)

View project on GitLab: ai_context

AI translate

The AI translate module is undergoing a strategic refactoring to support multiple ways of triggering translations. Currently, translations can only be requested through the user interface or command-line tools. Work is now underway to prepare the module to support a third method: agent-driven translation, where AI agents can perform translations automatically through a tool interface.

This expansion requires restructuring the module's core logic. Rather than maintaining separate code for each entry point, a unified service layer is being developed to handle all translation workflows. A draft merge request for refactoring entity translation into a shared orchestrator service has been opened, introducing new shared components that consolidate duplicated logic from existing interfaces.

This work positions the module to support more automated and programmatic translation workflows, which could reduce manual effort and enable new integration possibilities. The module remains in active development as the refactoring continues.

How can I help on this project?

  • Consider allocating additional development resources to accelerate the refactoring work, which will unlock new automated translation capabilities.
  • Identify early adopter teams who could benefit from agent-driven translation once available, to inform prioritization and testing.

Contributors: Sven Decabooter (svendecabooter)

View project on GitLab: ai_translate

Pinecone VDB Provider

The Pinecone VDB Provider is an integration module that connects Drupal to Pinecone's vector database service, enabling advanced search and content discovery capabilities powered by AI.

During this period, a new feature request was opened to add batching support for indexing operations. Batching allows the system to process multiple content items together rather than one at a time, which can significantly improve performance when indexing large volumes of content. This enhancement was inspired by similar work completed in a related database provider. The feature is currently awaiting review, with no development work started yet.

No code changes were merged or committed during this 24-hour window. The module continues in its current state while the batching feature is being considered.

How can I help on this project?

  • Clarify whether improved indexing performance is a priority for your content strategy, which would help determine resource allocation for the batching feature.
  • Consider connecting development teams across related database providers to share implementation approaches and accelerate delivery.

View project on GitLab: ai_vdb_provider_pinecone

AG-UI

AG-UI saw two new feature requests opened during this period, both focused on improving observability and troubleshooting capabilities. The first request concerns Add events for failures, which would enable the system to detect and respond to connection issues or remote service exceptions in a way that provides meaningful feedback to users. The second request, Debug should include timestamps, would add human-readable timing information to debugging output to help diagnose performance and sequencing issues.

Both items represent incremental improvements to the module's reliability and developer experience rather than core functionality changes. No code was merged or committed during this period. As these are enhancement requests for an early-stage module, the open items do not present immediate concerns.

How can I help on this project?

  • Clarify priority between reliability improvements and new feature development to guide resource allocation.
  • Ensure developer capacity is available to address observability requests if user feedback indicates troubleshooting is a barrier to adoption.

View project on GitLab: agui

References

[1] Document FunctionalJavascript test conventions and add CI check for ungrouped/unrecorded tests (ai, Issue)

[2] Fix PHPStan issues on API Explorer (ai, Issue)

[3] Batch embeddings for improved indexing performance (ai, Issue)

[4] Add an AI reranking processor for Search API (owned by the AI module) (ai, Issue)

[5] Add a configure-ai-settings agent skill for general AI settings (default models, providers, timeout, trusted domains) (ai, Issue)

[7] AI CKEditor tools (Tone, Summarise) fail with "Call to a member function get() on null" on Drupal 11.4.0 (ai, Issue)

[8] [1.x] Add a setup-automators agent skill to configure automators on fields (ai, Issue)

[9] AI CKEditor: Hardcoded “summarize” message shown for all plugins when no text is selected (ai, Issue)

[13] Enforce execution principal in AiAssistantApiRunner (no fallback to session user when executor is provided) (ai, Issue)

[15] Issue #3586543: Add AI reranking processor for Search API. (ai, MR)

[16] task: #3586562 Add a configure-ai-settings agent skill for general AI settings... (ai, MR)

[17] task: #3586563 [1.x] Add a setup-automators agent skill to configure automators on fields (ai, MR)

[23] feat: #3575346 AI CKEditor: Hardcoded “summarize” message shown for all... (ai, MR)

[24] feat: #3575346 AI CKEditor: Hardcoded “summarize” message shown for all... (ai, MR)

[25] feat: #3575346 AI CKEditor: Hardcoded “summarize” message shown for all... (ai, MR)

[26] replace hardcoded "summarize" no-selection message with overridable... (ai, MR)

[28] feat: #3575346 AI CKEditor: Hardcoded “summarize” message shown for all... (ai, MR)

[30] replace hardcoded "summarize" no-selection message with overridable getNoSelectedTextMessage() per plugin (ai, MR)

[32] Resolve #3574723 "Enforce execution principal" (ai, MR)

[33] Create a configurable search api processor for XB data (experience_builder, MR)

[35] Add AI ECA Interceptor submodule: intercept AI requests/responses through ECA (ai_integration_eca, Issue)

[36] Draft: Resolve #3584407 "Ai eca interceptor" (ai_integration_eca, MR)

[38] Issue #3558801: Route the OpenAI chat operation through the Responses API... (ai_provider_openai, MR)

[46] force_value is ignored when hide_property is unchecked (AI Agents 1.2.x) (ai_agents, Issue)

[49] fix: #3586045 persist force_value tool restrictions saved from the agent form (ai_agents, MR)

[61] Deleting optional props from an SDC/code component causes validation errors in config-defined component instances: regression for monolingual sites, bug for multilingual sites (canvas, Issue)

[63] Rudimentary conflict resolution UI (canvas, Issue)

[64] Introduce the authentication and security model for headless draft preview (canvas, Issue)

[71] Canvas AI: Client-side orchestration loop roadmap (canvas, Issue)

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

[96] chore(Internal HTTP API): #3591669 Heuristics for the content-entity-reference selection UI (canvas, MR)

[102] Draft: feat: Introduce the authentication and security model for headless draft preview (canvas, MR)

[103] chore(Component sources): #3591727 Store each prop's translatability per component version so deleting an optional prop keeps config-defined component trees valid (canvas, MR)

[105] feat(conflict): Add conflict resolution UI with viewport selection and route integration (canvas, MR)

[110] feat(Conflict detection): #3591601 "Conflict resolution via API call and layout endpoint support for published entities" (canvas, MR)

[113] Issue #3591663: Fail validation for props that do not exist on a component (canvas, MR)

[122] Clean up CCC form and table accessibility cleanup (ai_context, Issue)

[123] Clean up CCC documentation and API consistency (ai_context, Issue)

[125] AiContextAgentSettingsTest fails with ai_agents 1.3.2+ due to hardcoded edit form path (ai_context, Issue)

[133] Usage record page leaks context item content and grants links without access checks (ai_context, Issue)

[141] Issue #3586299: Fix test and documentation consistency before rc1. (ai_context, MR)

[143] Issue #3586304: Clean up CCC form and table accessibility cleanup (ai_context, MR)

[144] Issue #3586305: Fix context prefix default alignment and stale getManageUrl docs. (ai_context, MR)

[145] Issue #3586315: Avoid full catalog scan when agent has no scope subscriptions. (ai_context, MR)

[148] Issue #3581955: Add EventSubscriber to inject context into ai_ckeditor tools. (ai_context, MR)

[155] Issue #3586311: Harden usage record and listing access checks. (ai_context, MR)

[156] Issue #3586323: Update agent edit form path in test. (ai_context, MR)

[157] Refactor translation workflow to support Tool API and agent-driven translation (ai_translate, Issue)

[158] Draft: Refactor entity translation into a shared orchestrator service (ai_translate, MR)

[159] Draft: Refactor entity translation into a shared orchestrator service (ai_translate, MR)

Period: 7 July 2026 to 8 July 2026  |  Generated: 8 July 2026 05:00 GMT

To view the full AI prompts used to generate this digest check this file: 24h_2026-07-07_prompts.txt

Looking for older digests? Browse all digests here.