Daily Digest

Daily Digest – 2 August 2026

html

2026 Capabilities Progress

  1. Page generation — Progress continues on Canvas AI's agent and tools infrastructure, with a merge request enabling administrators to select which AI agent and tools power the chat interface [25][29]. An additional fix ensures AI-generated components fall back gracefully when referencing non-existent regions [36].
  2. Context management — The Context Control Center focused on user experience refinements, including work to clarify scope configuration guidance and improve conditional subcontext behavior [38][39]. Demo content is being prepared to illustrate how context items work in realistic scenarios [40].
  3. Background agents — Planning advanced on enabling ECA workflows to invoke Orchestration services, which would allow event-driven automation to trigger AI services without manual prompting [42]. No code was in progress; work remains in design phase.
  4. Design system integration — A folder assignment bug affecting newly created code components was fixed, ensuring components are correctly organized in Canvas libraries [27][35]. Accessibility issues in the editor interface were identified but not yet remediated [28].
  5. Content creation and discovery — Interactive refinement capability for field widgets is under development [5], and AI Automator token replacement functionality advanced with a merge request under review [1][6]. Platform compatibility work ensures core AI functionality remains stable [4][7][8].
  6. Advanced governance — Significant planning activity established a comprehensive EU AI Act compliance module (ai_disclosure) with fourteen detailed implementation issues covering disclosure tracking, editorial workflows, audit trails, and approval flows [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24]. No code in progress; work is in planning stage with legal compliance deadline of August 2026.
  7. Intelligent website improvements — No direct progress today.
  8. Multi-channel campaigns — Translation capability improvements continued with a fix for media entity language handling under review [41]. JSON Schema work for orchestration services would improve multi-service workflow composition but remains in planning [43].

Shipped

  1. AI API explorer compatibility fix merged — The API explorer now works correctly with both current and legacy provider implementations, maintaining compatibility across system versions [8].
  2. Canvas folder organisation bug resolved — Newly created code components now save to the correct folder selected by users, improving library organisation for site builders [27][35].
  3. Canvas AI region fallback protection in review — A merge request proposes fallback to the first available region when AI names a region that does not exist, preventing errors during automated component placement [36].
  4. Media translation language handling fix submitted — A merge request addresses invalid language code issues during media entity translation [41].

Ongoing

  1. Symfony AI Platform compatibility upgrade in progress — Work is underway to upgrade compatibility with Symfony AI Platform versions 0.8 through 0.12, addressing critical routing and provider interface issues that affect core functionality [4][7].
  2. EU AI Act compliance module in planning — A comprehensive transparency disclosure module is being designed to meet Article 50 requirements, with fourteen implementation issues covering field architecture, editorial interface, display formatting, and integration points for existing AI tools [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24].
  3. Canvas AI agent settings interface under review — A merge request allows administrators to explicitly choose which agent and tools are available in the chat interface, simplifying the current automatic orchestrator model [25][29].
  4. AI token stripping capability under review — A feature allowing administrators to remove unresolved placeholder tokens from AI prompts before submission is awaiting approval, which will reduce costs and improve response quality [1][6].
  5. Canvas accessibility violations identified — Six accessibility violations including two critical issues affecting keyboard navigation and colour contrast in the editor toolbar have been documented but remediation has not yet begun [28].

AI (Artificial Intelligence)

The AI module saw focused activity over the past 24 hours, primarily centered on compatibility updates and user experience planning.

A merge request was opened to upgrade the module's compatibility with newer versions of the underlying Symfony AI Platform library (versions 0.8 through 0.12). This work addresses critical compatibility issues that would prevent the system from functioning correctly, including routing and provider interface mismatches. Separately, a fix was merged to ensure the API explorer works correctly with both current and legacy provider implementations.

On the feature development side, work continued on allowing administrators to strip unresolved placeholder tokens from AI prompts before submission. This capability helps reduce unnecessary text being sent to AI services, which can impact both cost and response quality. A merge request for this functionality is now under review.

Two new user experience initiatives were opened for planning: improving the interface for Guardrails (the security and moderation controls that protect AI interactions) and reviewing the setup flow for MCP clients. Both are in early assessment stages.

How can I help on this project?

Consider prioritizing resources for the Symfony platform compatibility work, as this affects core functionality. Additionally, the two UX review initiatives (Guardrails and MCP clients) could benefit from dedicated design or product management time to move from planning to implementation.

Contributors: Dezső Biczó (mxr576), Rodel Duterte (rduterte)

View project on GitLab: ai

Drupal AI Initiative

Significant planning activity focused on EU AI Act compliance. A comprehensive transparency module specification was opened, responding to Article 50 requirements that took effect 2 August 2026. The regulation mandates visible disclosure of AI-generated content published to inform the public, with penalties up to EUR 15M or 3% of worldwide turnover for non-compliance.

The proposed ai_disclosure module would provide site operators with a structured disclosure system: a graduated scale of AI involvement (from metadata-only assistance to fully AI-generated content), reusable disclosure profiles to avoid duplicating text across content, and configurable defaults per content type. Content editors would see AI assistance suggestions when tools like translation, alt-text generation, or CKEditor assistance are used, with the ability to accept or dismiss them. The module would render EU-mandated icons and disclosure statements where required, include machine-readable metadata for platforms and search engines, and provide administrators with a compliance audit report.

Fourteen detailed implementation issues were opened covering the field architecture, editorial interface, display formatting, integration points for existing AI tools, multilingual support, and documentation. No code was merged during this period; work remains in the planning stage.

How can I help on this project?

  • Confirm organisational priority for EU AI Act compliance work and communicate target timeline for alpha release to align contributor effort.
  • Consider legal review of the disclosure framework design to validate alignment with Article 50 interpretation before implementation begins.

View project on GitLab: ai_initiative

Drupal Canvas

Over the past 24 hours the team fixed a folder organisation bug and continued work on two strategic priorities: evolving the AI chat interface and improving accessibility.

A fix for Incorrect folder assignment for newly created code components was merged, ensuring newly created code components are now saved to the folder users select rather than appearing in the wrong location. This directly improves the library organisation experience for site builders.

Two related issues advanced the roadmap to simplify how Canvas AI works. Canvas AI: Add an Agents & Tools settings form to select the main agent and chat Tools now has an open merge request that will allow administrators to explicitly choose which agent and tools are available in the chat interface, moving away from the current automatic orchestrator model. Separately, Canvas AI: Add Tools dropdown to chat UI is progressing the front-end counterpart, giving users a visible dropdown to select tools for each message.

A new accessibility audit identified 6 axe-core accessibility violations (2 critical) in the editor chrome, including keyboard navigation and colour contrast issues in the editor toolbar and controls. Remediation work has not yet begun.

How can I help on this project?

  • Assign or fund developer time to address the two critical accessibility violations in the editor; these affect every site builder using Canvas.
  • Review and approve the AI agent settings and chat UI work to unlock the simplified workflow outlined in the roadmap.

Contributors: **** (project_19391_bot_cb443e60680723bba08a5351bf595251), Ben Mullins (bnjmnm) [1], Lionel Enkaoua (heyyo), Rajab Natshah (RajabNatshah), Tekla Aivazashvili (aivazashvilit)

View project on GitLab: canvas

Context Control Center (CCC)

Activity this period focused on planning improvements to the user experience and scope configuration. Work continued on three issues addressing how the module presents guidance to editors and how conditional subcontext settings behave when disabled.

Progress was made on Decouple context item form descriptions from scope plugin descriptions, which addresses confusion caused by reusing the same guidance text in different parts of the interface. This will help editors understand scope settings more clearly in each context where they appear.

Planning also advanced on Improve conditional subcontext disabled-state behavior and admin UX, which will clarify what happens when the conditional subcontext feature is turned off and ensure editors understand why certain items become inactive.

Work continued on Prepare demo content and demo context items for the CCC to show how it works, which will provide realistic sample data to help new users understand how the module's components work together before configuring their own implementation.

How can I help on this project?

  • Confirm resourcing for user experience refinements to support a stable release.
  • Identify stakeholders who can review and validate the planned demo content to ensure it reflects realistic organisational use cases.

View project on GitLab: ai_context

AI translate

A fix addressing translation issues with media entities was submitted during the reporting period. The work targets a problem where media content was being processed with an invalid language code during translation operations, which would prevent these assets from being properly translated. A merge request addressing translation language handling for media entities is now under review.

This fix is important for organizations using the module to translate media-rich content, as it extends reliable translation capabilities beyond text to images, documents, and other media assets.

How can I help on this project?

  • Prioritize review and testing resources to evaluate the media translation fix currently awaiting merge.
  • Consider allocating development capacity to expand test coverage for media translation scenarios to prevent similar issues.
Contributors: szato (odi)

View project on GitLab: ai_translate

Orchestration

Activity this period focused on planning and feature design rather than code delivery. Two issues saw updates that clarify the module's strategic direction.

Work continued on adding an ECA action to invoke Orchestration services. This feature will enable workflows built in ECA (a popular rule engine) to call any service registered with the Orchestration module. Currently, the integration works in only one direction: external systems can trigger ECA workflows via Orchestration's HTTP interface. Completing the reverse path will allow organisations to use ECA's event and condition logic to trigger orchestrated services, expanding automation possibilities without custom code.

Separately, discussion advanced on using JSON Schema to define service inputs, outputs, and configuration. This architectural decision would establish formal contracts for each service, improving validation, documentation, and the ability to auto-generate user interfaces. The approach supports safer composition of complex workflows and better runtime error handling.

No code was merged this period. Both efforts remain in the planning and acceptance phase.

How can I help on this project?

  • Prioritise resourcing to move accepted feature work from planning into implementation, particularly the ECA integration which has clear user demand.
  • Support architectural decisions around JSON Schema adoption to unblock design work and maintain momentum on the module's roadmap.

View project on GitLab: orchestration

References

[1] AI Automator Token Replacement with removing unresolved Tokens (Checkbox?) (ai, Issue)

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

[5] Add interactive refinement capability to Field Widget Actions (ai, Issue)

[6] Issue #3586535: AI Automators strip unresolved tokens from prompts (ai, MR)

[7] feat: Upgrade to Symfony AI Platfrom 0.12 (ai, MR)

[8] fix: #3586470 Make API explorer work with Symfony platform and legacy providers (ai, MR)

[9] [Meta] ai_disclosure: EU AI Act Article 50 transparency module (ai_initiative, Issue)

[10] ai_image_alt_text companion: suggest ai_metadata via the ai_disclosure recorder (ai_initiative, Issue)

[11] ai_disclosure: DisclosureRecorder public API, suggestion storage and Accept/Dismiss flow (ai_initiative, Issue)

[12] ai_translate companion: apply ai_translated grade via the ai_disclosure recorder (ai_initiative, Issue)

[13] ai_ckeditor companion: suggest ai_partly_assisted via the ai_disclosure recorder (ai_initiative, Issue)

[14] ai_automators companion: report AI field writes to the ai_disclosure recorder (ai_initiative, Issue)

[15] ai_disclosure: Documentation - module docs and recorder API integrator guide (ai_initiative, Issue)

[16] ai_disclosure: Multilingual and revision hardening with test coverage (ai_initiative, Issue)

[17] ai_disclosure: Field type, DisclosureResolver service and inherit storage model (ai_initiative, Issue)

[18] ai_disclosure: Compliance report at /admin/reports/ai-disclosure (ai_initiative, Issue)

[19] ai_disclosure: Machine-readable marking - JSON-LD and meta output with IPTC digitalSourceType (ai_initiative, Issue)

[20] ai_disclosure: Formatter theming - EU icons, disclosure template, policy settings (ai_initiative, Issue)

[21] ai_disclosure: Widget UX - sidebar group, mode states, computed status line, mdxeditor (ai_initiative, Issue)

[22] ai_disclosure: Disclosure profile config entity and admin UI (ai_initiative, Issue)

[23] ai_disclosure: Grade config entity with shipped defaults and admin UI (ai_initiative, Issue)

[24] ai_disclosure: Scaffold module - info, permissions, settings shell, EU icon assets (ai_initiative, Issue)

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

[27] Incorrect folder assignment for newly created code components (canvas, Issue)

[28] Canvas editor: 6 axe-core accessibility violations (2 critical) in the editor chrome — topbar menubar, unnamed zoom button, region overlays, badge contrast, viewport, landmarks (canvas, Issue)

[29] Issue #3591882: Add an Agents & Tools settings form. (canvas, MR)

[35] #3591828 Refetch folder list after newly exposing a code component (canvas, MR)

[36] fix: #3591876 Fall back to first region when the AI names a region that does not exist (canvas, MR)

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

[39] Improve conditional subcontext disabled-state behavior and admin UX (ai_context, Issue)

[40] Prepare demo content and demo context items for the CCC to show how it works (ai_context, Issue)

[41] fix: #3585532 Translation language (und) when translating media entities. (ai_translate, MR)

[42] [orchestration_eca] Add "Execute Orchestration service" ECA action to allow ECA to invoke registered services (orchestration, Issue)

[43] Using JSON Schema to define "input", "output" and "config" (orchestration, Issue)

Period: 2 August 2026 to 3 August 2026  |  Generated: 3 August 2026 08:38 GMT

To view the full AI prompts used to generate this digest check this file: 24h_2026-08-02_prompts.txt

Looking for older digests? Browse all digests here.