Daily Digest

Daily Digest – 10 August 2026

2026 Capabilities Progress

  1. Page generation — Progress was made fixing a critical bug where AI-generated pages fail completely when an invalid region name is specified [153], and work continues on the AI page builder agent and component placement tools [122][129]. Next milestone is completing the client-side orchestration loop and ensuring robust error handling.
  2. Context management — Significant progress toward the 1.0 stable release: the token budget system was simplified by removing the redundant max_items limit [201], priority-based selection ordering was implemented [198], and documentation for roles and permissions was completed [205]. Next milestone is completing QA testing before final release.
  3. Background agents — Planning continues on the architectural challenge of supporting sessionless, queued, and scheduled AI execution [12], but no code progress occurred today. This remains a blocker for automated content generation and scheduled workflows.
  4. Design system integration — Work advanced on Canvas AI integration with fixes for component placement validation [153] and improvements to the props exposure system [123], while the code editor now shows better previews of content-entity-reference data [162]. Next milestone is completing the client-side orchestration and tools dropdown UI [151].
  5. Content creation and discovery — The AI CKEditor integration saw a fix for Gemini 2.x model compatibility [10], and search relevance improved with a reranker fix prioritizing chunk text over indexed fields [8]. A critical content loss bug in text generation automators remains under investigation [11].
  6. Advanced governance — Canvas made progress on conflict resolution with autosave hash stabilization work [173] and workspace staging exploration [134], while audit capabilities were strengthened through cache tag inheritance fixes [169] and translation synchronization improvements [174].
  7. Intelligent website improvements — Feedback collection capability was added to the Langfuse observability module [223], enabling sites to capture user ratings that can inform AI learning and optimization. The thumbs-up/down icon set was added to support feedback interfaces [9].
  8. Multi-channel campaigns — Substantial progress on AI Translation toward the DrupalCon Rotterdam demo [218]: work began on translation result caching and deduplication [208], automatic translation of referenced taxonomy terms [214], and refactoring to support tool-based and agent-driven workflows [213].

Shipped

  1. AI Initiative partner outreach communications completed — More than 30 partnership conversations were held with organizations including Calibrate, Oomph, Wunder, Pantheon, EPAM, and others, with communication tracking issues closed to record completed calls and outreach [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60].
  2. Tool API execution contract and validation hardened — Execution interfaces were clarified, input validation now covers all items instead of just the first, error messages became translatable, and AI integration was strengthened to handle common formatting variations automatically [104][105][106][107][108][109][110][111][112][113][114][115][116][117][118][119][120][121].
  3. Context Control Center simplified configuration and documented permissions — Administrators now configure only token budgets rather than both item counts and budgets, reducing complexity while maintaining control, and role assignment guidance was published [201][205].
  4. Canvas translation validation and cache handling fixed — Publishing symmetrically-translated pages with non-translatable component edits now validates correctly, and array property shapes properly inherit cache tags to ensure invalidation works as expected [174][169].
  5. Langfuse feedback collection capability released — A new submodule enables sites to capture and score user responses to AI features, supporting measurement of satisfaction and response quality in production [223].

Ongoing

  1. AI module prompt rendering and content generation defects under review — HTML escaping corrupts context before reaching providers, text generation silently drops content with certain field configurations, and Gemini 2.x model compatibility issues remain under investigation [4][11][10][14].
  2. AI Agents content creation and interface improvements in review — Seven merge requests address data handling bugs preventing successful article creation, schema editor upgrades, and field information leaks, with all awaiting final review [69][70][71][72][73][74][75][76][77][78].
  3. Canvas AI page builder region validation under development — When AI assistants specify non-existent regions, the entire page build is discarded rather than validated before placement, with a fix proposed but not yet in progress [124][153].
  4. Context Control Center selection behavior and documentation nearing 1.0 — Work continues on scoring alignment, token count persistence for budget visibility, and documentation covering real-world use cases and integration patterns [184][192][188][189][198][199][200][204][206][207].
  5. AI Translate demonstration and cost-saving features in development — DrupalCon Rotterdam demo requires one-click translation from editing interface and multilingual search, with caching, deduplication, and automatic reference translation under active development [218][208][214].

AI (Artificial Intelligence)

The AI module saw continued community coordination and issue triage over the past 24 hours. A contribution meeting was held to align the community on priorities and discuss active work. Several new defects and integration requests were identified and are now being addressed.

Key issues moving forward include a prompt rendering bug where HTML content is unintentionally escaped before being sent to AI providers, and a content loss issue affecting text generation workflows when using certain field configurations. A UI error preventing tool deletion has reached ready-to-be-committed status. Work continues on embedding support via Symfony AI and test infrastructure documentation to improve contributor onboarding. An incompatibility with newer Gemini models has been documented and is under investigation.

The plan to support background AI execution remains an open architectural challenge that will unblock use cases like scheduled content generation and queue-based processing.

How can I help on this project?

  • Champion a decision on the background execution architecture to unblock queued and scheduled AI workflows.
  • Consider resourcing for compatibility testing across major AI providers (OpenAI, Gemini, Anthropic) to reduce integration defects.

Contributors: Ann Mary Sruthy (annmarysruthy), Rob Loach (robloach), Scott Euser (scotteuser), Shivam Sen (shivamsen_12579)

View project on GitLab: ai

Drupal AI Initiative

Activity over the past 24 hours focused on two main areas: partnership outreach and planning for the upcoming DrupalCon Rotterdam demonstration.

Significant effort was invested in expanding the initiative's partner base. More than 30 outreach conversations were documented and closed, resulting in commitments from multiple organizations including Calibrate, Oomph, Specbee, Cambridge Press, Reading Room, Wunder, Mearra, Drupalforge, E-sepia, Integral Vision, Brainsum, Optasy, Drunomics, Kalamuna, Pantheon, Esinergia, Foster Interactive, EPAM, iTech4Web, Pronovix, and Zivtech. Several additional organizations are still considering participation. This partnership growth directly supports the initiative's funding and collaborative development capacity.

Planning continued for the Rotterdam demonstration, which aims to showcase multiple AI capabilities working together in a single site that attendees can deploy themselves. The UX team met to discuss content requirements, tone guidelines, and compliance with EU AI Act disclosure requirements. Work is also underway to update partner badges to reflect the collaborative "Partner" terminology rather than "Sponsor."

How can I help on this project?

  • Ensure design resources are allocated to complete the partner badge updates before Rotterdam
  • Support partnership follow-up with organizations still considering participation to secure additional funding commitments
  • Help prioritize demo content creation to meet the Rotterdam timeline

View project on GitLab: ai_initiative

Model Context Protocol

The Model Context Protocol module saw limited activity during this period, with two authentication-related issues receiving updates. Both items involve improving how the module handles secure connections between clients and servers.

Work continues on Token Auth is not working, which addresses problems preventing administrators from enabling token-based authentication through the configuration interface. This issue affects the usability of security features that control how external systems connect to Drupal.

Additionally, Add OAuth Scope Validation for MCP Authentication remains under development. This work aims to give administrators better control over what permissions are granted when external clients authenticate, ensuring that only necessary access rights are approved rather than automatically accepting all requests.

No code was merged or committed during this period. Both issues are assigned and part of ongoing sprint work, indicating steady progress within the AI Initiative's development roadmap.

How can I help on this project?

  • Ensure the team has sufficient review capacity to evaluate and merge authentication fixes once development is complete.
  • Clarify security policy requirements for OAuth scope handling to guide technical implementation decisions.

View project on GitLab: mcp

AI Agents

The AI Agents module saw significant development activity focused on fixing defects and improving the authoring experience. Seven new merge requests were opened addressing issues ranging from data handling bugs to interface improvements.

Work is underway to fix a bug where the content creation tool was storing field values incorrectly, which prevented AI assistants from successfully creating articles and other content. Separately, improvements are being made to the schema editor used when configuring agents, replacing a basic text box with a specialized code editor that validates JSON syntax and highlights errors as administrators work.

Additional merge requests target bugs in field information handling, tool status reporting, and support for newer AI model series. A regression test was added to ensure agents remain immediately usable after creation without manual cache clearing.

One exploration issue was closed during the period: work investigating how to export and synchronize Drupal-defined agents to external AI platforms like Claude Code. This signals a decision point on integration strategy with third-party agent ecosystems.

All open merge requests are in review. The module continues to mature its content creation capabilities and administrative tooling.

How can I help on this project?

  • Allocate reviewer capacity to accelerate the seven open merge requests, particularly the content creation bug fix which blocks practical use cases.
  • Clarify strategic direction on external platform integration following closure of the export/sync exploration work.

Contributors: Ahmad Khader (Ahmad-Khader), Bharat Kelotra (bharatkelotra), Joshua Fernandes (joshua1234511), Prabhavathi Vanipenta (prabha1997)

View project on GitLab: ai_agents

Tool API

The Tool API module made significant progress toward its RC1 release during this period, with 18 issues closed and 17 merge requests integrated. The work focused on hardening the plugin contract and input/output system ahead of the interface freeze.

A major area of activity was fixing validation and input handling bugs that could have caused tools to fail in production. Multiple inputs now validate all items instead of just the first one, and list-type inputs are now checked properly. Input values that cannot be stored now produce clear error messages instead of generic exceptions. Error messages throughout the system are now translatable and no longer expose raw technical details to end users.

The execution contract was reworked to use clearer naming and properly typed interfaces. An unimplemented caching interface was removed rather than shipped incomplete. Several identifiers were prefixed to follow Drupal naming conventions and avoid namespace collisions with other modules.

Integration with AI systems was strengthened: the connector now handles tools with unmet requirements gracefully, input schemas now include richer constraint information, and the system automatically corrects common formatting variations that AI models produce (such as sending JSON numbers instead of strings for decimal inputs, or wrapping single values in arrays).

Comprehensive test coverage was added for the typed input and output system, configuration precedence, and validation flows. One open issue tracks a schema conversion bug affecting tools that use bare regex patterns.

How can I help on this project?

  • Confirm timeline and scope expectations for the RC1 release to ensure the remaining schema bug can be resolved without delaying the freeze.
  • Support documentation resourcing: project page automation is ready but needs enablement, and tool authoring guidance would benefit from dedicated technical writing.

Contributors: Ajit Shinde (ajits), Alex Urevick-Ackelsberg (alexua) [2], Michael Lander (michaellander) [15]

View project on GitLab: tool

Drupal Canvas

Drupal Canvas made progress across data integrity, translation workflows, and AI capabilities during this reporting period.

A field configuration conflict affecting test modules on the standard profile was resolved, improving compatibility for developers manually testing content-entity-reference features. A fix for array prop shape cache tag inheritance was merged, ensuring components with array-based properties (such as multi-value image fields) correctly invalidate when underlying configuration changes. Work advanced on normalizing autosave hash generation, consolidating duplicated typed-data logic to strengthen conflict detection reliability.

Two translation-related issues moved forward: a publishing validation bug affecting symmetrically-translated pages was merged, and discussion continued on aligning Canvas translation operations with core content_translation and TMGMT. An AI page-builder issue where incorrect region names abort entire builds is under review, with a fix proposed to validate regions before placement.

A header region height calculation bug introduced in 1.9.0 was reported, blocking editor interaction on some sites. Work on editable stored patterns was merged, allowing users to rename and modify patterns after creation. A cache context tagging error in content template normalization was identified and is being corrected.

How can I help on this project?

Contributors: **** (project_19391_bot_cb443e60680723bba08a5351bf595251), Adam G-H (phenaproxima), Alamgir Hossain (alamgir.hossain.cse), Ben Mullins (bnjmnm), Christian López Espínola (penyaskito) [1], Dave Long (longwave), Feliksas Mazeikis (f.mazeikis), Fons Vandamme (f0ns), Kunal Sachdev (kunal.sachdev), Lauri Timmanee (lauriii), Nodar Davituri (Davituri), Rajab Natshah (RajabNatshah), Ted Bowman (tedbow), Tim Plunkett (tim.plunkett) [1], Wim Leers (wimleers)

View project on GitLab: canvas

Context Control Center (CCC)

The project advanced toward its 1.0 stable release during this period with two significant changes merged. The first simplifies how the system limits the amount of information sent to AI agents. Previously, site administrators configured both a maximum number of context items and a token budget (a measure of content size). This created unnecessary complexity, since context items vary widely in length. The change to Remove max_items context limit and rely on the token budget means administrators now configure only the token budget, providing more accurate control over what reaches AI agents while reducing setup complexity.

The second merged change provides documentation for the role and permission matrix, giving site builders clear guidance on how to assign permissions to different types of users—from AI feature consumers to content authors to system administrators.

Work continues on several fronts preparing for stable release. Progress is underway on improving how the system selects and orders context when multiple items compete for limited space, persisting token counts to give editors better visibility into budget usage, and aligning scoring behavior to make selection more predictable. Documentation covering real-world use cases and integration patterns is also in development.

How can I help on this project?

Support resourcing for the QA testing phase, which requires test scripts and coordination before final release. Consider whether roadmap decisions are needed around post-1.0 features currently proposed for future sprints. Ensure documentation and use-case materials receive stakeholder review to confirm they address real deployment scenarios.

Contributors: Abhinav Jha (ajabhinavjha30), Kristen Pol (kepol) [2]

View project on GitLab: ai_context

AI Dashboard

The AI Dashboard module saw focused maintenance work during this period. A visual refinement was completed to ensure consistent display of descriptive text throughout the interface, improving overall readability and user experience.

This represents steady progress on polish and usability as the module matures. No major risks or blockers were identified during this period.

How can I help on this project?

  • Consider whether the project needs dedicated design or user experience support to identify further usability improvements.
  • Assess whether additional testing resources could help accelerate the path to a stable release.

Contributors: Rob Loach [1]

View project on GitLab: ai_dashboard

AI translate

The AI translate project saw significant planning activity over the past 24 hours, with fourteen issues updated as teams prepare for a demonstration at DrupalCon Rotterdam 2026. Work is underway to make the module a production-ready alternative to manual content translation, with particular focus on cost reduction, editor experience, and integration with multilingual search.

Key areas of progress include:

Several bugs affecting media translation and layout builder workflows remain open and postponed, which is typical for a module at this stage.

How can I help on this project?

Confirm resource allocation for the DrupalCon Rotterdam demo delivery. Clarify whether automatic translation of referenced entities should be opt-in or default behaviour to unblock the design decision. Consider funding a dedicated QA cycle for the postponed media and workflow bugs once core features stabilise.

Contributors: Christoph Breidert (breidert)

View project on GitLab: ai_translate

FlowDrop

During the past 24 hours, FlowDrop reached a significant quality milestone by strengthening its internal code analysis to the highest available standard. The project advanced through two major quality levels in succession, improving type safety throughout the platform. This work touched every component of the system, including workflows, pipelines, sessions, jobs, memory systems, chat functionality, and the orchestration layer.

These improvements reduce the likelihood of runtime errors and make the codebase more maintainable over time. The changes ensure that data flowing through the system is more predictable and easier to validate, which should result in more reliable behaviour for users. Additionally, spelling checks were reinstated and British English was adopted as the project's documentation standard, improving consistency across all written materials.

This type of foundational work does not deliver immediately visible features but positions the project for faster, safer development of new capabilities in the future.

How can I help on this project?

Consider clarifying the roadmap priorities for user-facing features now that code quality has been strengthened. If the project is approaching a release milestone, allocating time for user testing or documentation review could ensure a smooth launch.

Contributors: Shibin Das [60]

View project on GitLab: flowdrop

Langfuse

The Langfuse module made notable progress in the last 24 hours with the completion of work to enable user feedback collection. A new feedback capability was merged, adding a dedicated submodule that allows sites to capture and score user responses. This enhancement will enable organizations to gather structured feedback data, which can inform product improvements and better understand how users interact with AI-powered features.

The change represents a significant expansion of the module's functionality, moving beyond basic tracking to include mechanisms for measuring user satisfaction and response quality. This positions the module to support more comprehensive observability of AI features in production environments.

How can I help on this project?

Consider allocating resources for user testing of the new feedback features to validate real-world usability. Support outreach to early adopters who could provide case studies demonstrating business value. Evaluate whether additional investment is needed to accelerate progress toward a stable release.

Contributors: Abhisek Mazumdar (abhisekmazumdar) [1]

View project on GitLab: langfuse

AG-UI

No development activity was recorded during this period. One new performance issue was opened concerning chunkDelay/chunkPadding applied per-event compounds into real latency/timeout risk on long or tool-heavy runs. The report identifies that artificial delays and padding are being applied to every event sent to the browser, not just content chunks. For conversations involving multiple steps or tool usage, this overhead accumulates and can add several seconds of unnecessary delay to each response, which may affect user experience in production deployments.

The issue remains unassigned and no solution is yet in development. Given the module's maturity stage, open performance issues are expected as the project evolves.

How can I help on this project?

  • Allocate developer time to investigate and resolve the performance overhead issue, particularly if this module supports user-facing or time-sensitive workflows.
  • Clarify priority: decide whether optimising response times is critical for the current release or can be deferred to a later phase.

View project on GitLab: agui

References

[1] Add AIDmi to Other modules/recipes that work with the AI module (ai, Issue)

[2] Drupal AI Contribution meeting 2026-08-10 (ai, Issue)

[4] Twig autoescaping in AiPromptHelper::renderPrompt() corrupts {{ context }} and {{ raw_context }} before they reach the provider (ai, Issue)

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

[9] Add thumbs-up and thumbs-down icons to the shared AI icon set (ai, Issue)

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

[11] LLM: Text / LLM: Text (simple) automators silently drop Base Field content when Base Field and Target Field are both long-text types (Basic input mode) (ai, Issue)

[12] [PLAN] Propagate and enforce an explicit executor for sessionless, queued, and outside-in AI runs (ai, Issue)

[14] bug: #3586661 Twig autoescaping in AiPromptHelper::renderPrompt() corrupts {{... (ai, MR)

[15] #3582633: Add Functional JavaScript testing conventions and validation script (ai, MR)

[19] Communication Binarywork to join initiative (ai_initiative, Issue)

[20] Communication Genero inc to join initiative (ai_initiative, Issue)

[21] Communication Smile to join initiative (ai_initiative, Issue)

[22] Communication Evolvingweb to join initiative (ai_initiative, Issue)

[23] Communication Digital polygon to join initiative (ai_initiative, Issue)

[24] Communication Sparkfabrik to join initiative (ai_initiative, Issue)

[25] Communication Promet source to join initiative (ai_initiative, Issue)

[26] Communication Redhat to join initiative (ai_initiative, Issue)

[27] Communication Omedia to join initiative (ai_initiative, Issue)

[28] Communication Faktorial to join initiative (ai_initiative, Issue)

[29] Communication NTTdata to join initiative (ai_initiative, Issue)

[30] Communication Swis to join initiative (ai_initiative, Issue)

[31] Communication Hintogroup to join initiative (ai_initiative, Issue)

[32] Communication Mobomo to join initiative (ai_initiative, Issue)

[33] Communication Kadabra IT to join initiative (ai_initiative, Issue)

[34] Communication EPAM joining (ai_initiative, Issue)

[35] Communication iTech4Web (ai_initiative, Issue)

[36] Call Pronovix to join (ai_initiative, Issue)

[37] Call Zivtech to join Initiative (ai_initiative, Issue)

[38] Communication Joshi consulting (ai_initiative, Issue)

[39] Communication Civiactions (ai_initiative, Issue)

[40] Call IOgroup CEO (ai_initiative, Issue)

[41] Call 2 calibrate (ai_initiative, Issue)

[42] Call 1 calibrate (ai_initiative, Issue)

[43] Call with Oomph (ai_initiative, Issue)

[44] Communication with Specbee (ai_initiative, Issue)

[45] Call with Cambridge press (ai_initiative, Issue)

[46] Call with Reading room to join initiative (ai_initiative, Issue)

[47] Went to see Vesa from Wunder in Cascais to join AI initaitive (ai_initiative, Issue)

[48] Met with team Mearra in Athens (ai_initiative, Issue)

[49] Call 2 with Drupalforge (ai_initiative, Issue)

[50] Call 1 with drupalforge (ai_initiative, Issue)

[51] Call/mail with Yorgos E-sepia (ai_initiative, Issue)

[52] Call with integral vision (ai_initiative, Issue)

[53] Call with Brainsum (ai_initiative, Issue)

[54] Call with Optasy (ai_initiative, Issue)

[55] Call with wolfgang Drunomics (ai_initiative, Issue)

[56] Call with Kalamuna (ai_initiative, Issue)

[57] Call 2 with Josh from Pantheon (ai_initiative, Issue)

[58] Call 1 with Josh from Pantheon (ai_initiative, Issue)

[59] Call with Esinergia (ai_initiative, Issue)

[60] Call with Aidan Foster interactive (ai_initiative, Issue)

[69] ContentEntitySeeder stores multi-property field values as Array (ai_agents, MR)

[70] Issue#3586025:GetEntityFieldInformation-Leaks-Target_entity_type-Target_bundle... (ai_agents, MR)

[71] Issue (ai_agents, MR)

[72] Issue3569280:AI-Agents-Explorer-Chat-Generation-Does-Not-Support-Series-Models fixes (ai_agents, MR)

[73] Issue#3566126:Force-Value-Dropped-Tool-Property-Restrictions fixes. (ai_agents, MR)

[74] Issue #3560681: Add per-tool execution ordering constraints for agent loops. (ai_agents, MR)

[75] Issue#3565169:Schema-Error-AI-Agents-CKEditor-Plugin fixes. (ai_agents, MR)

[76] #3586042: Use the ai_json_schema editor for the structured output schema field. (ai_agents, MR)

[77] Issue #3586042 Use the `ai_json_schema` editor for the AI Agent structured-output schema field (ai_agents, MR)

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

[104] fix: #3582986 require the logger channel and document the subclass redeclaration contract (tool, MR)

[105] fix: #3582985 cast ArrayObject schema values to arrays before building AI tool properties (tool, MR)

[106] Fix fromDataDefinition: fall back for missing labels and toArray type keys. (tool, MR)

[107] Resolve "Coerce JSON numbers to strings for decimal inputs" (tool, MR)

[108] Enforce checkRequirements() in execute() and exclude unmet tools from the AI function list. (tool, MR)

[109] Add kernel coverage: locked/configuration inputs, refiner flow, definition... (tool, MR)

[110] fix: #3582964 own execution contract: result value interfaces, accessors on... (tool, MR)

[111] fix: #3582947 tool_ai_connector does not handle tools returning list of entities (tool, MR)

[112] fix: #3582962 correct API signatures: factory overrides, null context values,... (tool, MR)

[113] fix: #3582965 drop the unimplemented cacheability contract from ToolInterface (tool, MR)

[114] fix: #3582963 prefix the adapter alter hook and cache key; keep the generic... (tool, MR)

[115] fix: #3582982 validate list definition items and name failing positions (tool, MR)

[116] Make input violation messages translatable; log unexpected exceptions instead... (tool, MR)

[117] fix: #3582961 fail closed when tool input is invalid or missing (tool, MR)

[118] Rethrow un-storable input values as InputException naming the input. (tool, MR)

[119] fix: #3582960 validate every multiple input item (tool, MR)

[120] Extend input coercion: JSON-string structures, enum near-misses, datetime... (tool, MR)

[121] Resolve "Emit richer JSON Schema: map remaining constraints and type formats" (tool, MR)

[122] Canvas AI: Add a dev page builder agent and wire the placement tools (canvas, Issue)

[123] Canvas AI: Expose Props of Blocks to the Agent (canvas, Issue)

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

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

[134] Stage Canvas autosaves in a dedicated Workspace (canvas, Issue)

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

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

[162] Add a entity reference prop data shape preview to code editor (canvas, MR)

[169] #3591917 Array prop shapes inherit their item shape's cache tags (canvas, MR)

[173] refactor(Conflict resolution): #3591785 Strengthen AutoSaveManager to generate stable hashes (canvas, MR)

[174] fix: #3591820 Synchronize non-translatable component inputs before validating auto-save publish (canvas, MR)

[184] Align scope scoring, language service, and contextual selection behavior (ai_context, Issue)

[188] Document real-world use cases for CCC 1.0 (ai_context, Issue)

[189] Document how to use RAG with CCC 1.0 (ai_context, Issue)

[192] Persist context-item token counts and show token-budget usage in CCC UI (ai_context, Issue)

[198] Issue #3586376: Add context item priority field and selection tie-breakers. (ai_context, MR)

[199] Issue #3586392: Clarify and stabilize scope plugin extension boundary for 1.0. (ai_context, MR)

[200] Issue #3586388: Align scope scoring and contextual detection. (ai_context, MR)

[201] Issue #3586378: Remove max_items context limit and rely on token budget. (ai_context, MR)

[204] Issue #3586379: Persist token counts and show budget usage in CCC UI. (ai_context, MR)

[205] Issue #3586395: Document the AI Context role and permission matrix for 1.0. (ai_context, MR)

[206] Issue #3586375: Install target_entities storage when DER is enabled later. (ai_context, MR)

[207] Issue #3586383: Pre-1.0 context selection reliability and diagnostics follow-up. (ai_context, MR)

[208] Feature: Translation result caching and cross-field deduplication (ai_translate, Issue)

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

[214] Auto-translate referenced taxonomy terms (and other translatable entities) when translating a node (ai_translate, Issue)

[218] [META] AI Translation — DrupalCon Rotterdam 2026 demo (ai_translate, Issue)

[223] Issue #3548465: Add createScore() for feedback and a langfuse_feedback submodule (langfuse, MR)

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

To view the full AI prompts used to generate this digest check this file: 24h_2026-08-10_prompts.txt

Looking for older digests? Browse all digests here.