Pull Request Explorer

Summary

Improve workspace file explorer UX

Enhances the file explorer with new view modes, filters, and improved navigation, boosting developer productivity and reducing context switching.

Health Assessment

• • Fast review and single iteration indicate a smooth process with minimal friction.

Summary

Record device vs transport time for tools

Adds instrumentation to split device execution time from transport latency, enabling smarter decisions on where to run agent logic and improving performance insights.

Health Assessment

• • The PR was reviewed and merged within 8 hours, with a single commit and no back‑and‑forth, indicating a smooth, low‑friction process.

Summary

Move RecentItem to shared types

Centralizes RecentItem DTO for consistency across server and client, preserving runtime behavior.

Health Assessment

• • Quick refactor with minimal changes, merged within 7.4 hours, indicating a smooth process.

Summary

Refactor GitHub connector to use Market OAuth

Centralizes OAuth handling, improves security by routing GitHub requests through Market’s trusted proxy, reducing token exposure and simplifying authentication.

Health Assessment

• • The PR was reviewed and merged within 2.3 hours with a single AI‑generated review comment, indicating a smooth, low‑friction process and minimal rework.

Summary

Fix Codex resume writer race condition

Ensures reliable message sending after stopping a Codex run, preventing data loss and improving user experience.

Health Assessment

• • Rapid review with automated AI comments led to quick resolution; large code changes across backend and frontend; race condition fix improves stability.

Summary

Add TRAE provider binding support

Enables heterogeneous agents to bind to TRAE providers, expanding integration options and improving model management.

Health Assessment

• • The PR was reviewed quickly and required only a single round of changes, indicating efficient collaboration despite its large scope.

Summary

Add Elasticsearch full-text search provider

Adds a robust, multi-language full-text search capability using PostgreSQL and Elasticsearch, improving content discoverability and user experience.

Health Assessment

• • Large-scale addition of search provider with extensive iterations and AI-assisted reviews, indicating significant effort and potential integration complexity.

Summary

Record local hetero agent runs as snapshots

Adds durable execution snapshots for CLI agent runs, enabling unified view, crash recovery, and improved debugging for users and developers.

Health Assessment

• • The PR was merged quickly with a single commit and minimal review, indicating a smooth process despite its large scope.

Summary

Align agent search results on mobile

Improves mobile UI by aligning agent search results, ensuring consistent layout and better user experience.

Health Assessment

• • Quick review and single commit indicate low complexity and high confidence in the changes.

Summary

Fix Kimi Code model gating by profile

Ensures Kimi Code models are only offered when compatible, improving reliability and reducing errors for users.

Health Assessment

• • Quick fix with minimal rework; AI review helped catch issues early.

Summary

Make Todo Tray List Scrollable

Enables scrolling within the Todo tray, improving usability when many items are present.

Health Assessment

• • Quick fix with minimal changes, AI review helped catch issues early.

Summary

Fix Render Gallery label word wrap

Improves UI readability by preventing label text from breaking mid-word, ensuring consistent layout across panel widths.

Health Assessment

• • Merged within 45 minutes, minimal review, no tests added.

Summary

Fix renderer OTA main bundle hash

Ensures OTA updates correctly detect changes, preventing unnecessary updates and improving stability for desktop users.

Health Assessment

• • PR had a rapid initial review but required multiple iterations, indicating complex logic changes; large code churn suggests significant impact on desktop OTA functionality.

Summary

Graduate In-App Browser from Labs

All desktop users can now use the in-app browser without a Labs toggle, improving usability and reducing friction. Verification failures no longer trigger urgent inbox briefs.

Health Assessment

• • Merged within 9 hours with minimal rework, indicating a straightforward implementation and low risk.

Summary

Replace draft label with icon

Updates the UI to use a compact pencil icon for draft topics, improving visual clarity and accessibility for users.

Health Assessment

• • Quick, single-commit PR with minimal changes and fast merge, indicating low complexity and high confidence.

Summary

Fix openapi: relay Gemini events to Claude

Ensures Claude Code receives proper content from Gemini, preventing empty responses and improving reliability.

Health Assessment

• • Quick fix with minimal rework; AI review helped catch edge cases.

Summary

Fix notification spam for background runs

Stops unnecessary OS banner notifications from background tasks, bots, and eval runs, improving user experience.

Health Assessment

• • Quick, single-commit fix with minimal code churn and no review back‑and‑forth.

Summary

Fix workspace provider deep-link redirect

Ensures users remain within their workspace when accessing provider links, preventing accidental logout and improving navigation.

Health Assessment

• • Rapid iteration with AI-assisted review enabled a quick fix; moderate scope but low risk to production.

Summary

Fix desktop memory reporting accuracy

Improves accuracy of renderer memory metrics in the desktop app, enhancing performance monitoring for users.

Health Assessment

• • Fast review and merge cycle (14.6h) with minimal rework; single AI-assisted review comment and no major conflicts indicate a smooth process.

Summary

Hide runtime-only server models from heterogeneous agents

Prevents hidden runtime-only model aliases from being exposed to heterogeneous agents, ensuring accurate capability matrix and secure model selection.

Health Assessment

• • Quick fix with minimal changes and tests, merged within 36 minutes, indicating low complexity and high confidence.

Summary

Normalize heterogeneous review inputs

Ensures consistent handling of review payloads, preventing invalid data from causing errors and ghost states, improving reliability for agents.

Health Assessment

• • Fast cycle time and minimal review indicate a straightforward bug fix, but the large number of added lines suggests careful testing was required to maintain stability.

Summary

Fix OpenAPI Responses for Grok Build

Fixes a critical bug in Grok Build's OpenAPI responses, ensuring proper token usage data and preventing serialization errors. This improves reliability for users interacting with the build.

Health Assessment

• • Quick bug fix with minimal changes and no major rework, indicating a smooth review process.

Summary

Fix OpenAPI to accept Pi responses

Ensures Pi responses are correctly routed, preventing 502 errors and empty assistant content, improving reliability for users.

Health Assessment

• • Rapid merge with minimal changes and a regression test to ensure stability.

Summary

Fix AVIF file recognition as images

Enables desktop agents to handle AVIF images without conversion, improving media handling and reducing processing overhead.

Health Assessment

• • Rapid review and merge with minimal iterations indicates a straightforward bug fix. The AI review tool provided automated feedback, contributing to the fast turnaround.

Summary

Fix gateway init race for heterogeneous runs

Eliminates a race condition that caused missing approval cards during heterogeneous agent runs, ensuring reliable operation and a smoother user experience.

Health Assessment

• • The PR was reviewed and merged within 2.7 hours with a single review comment, indicating a smooth process. Despite the large code change (over 600 lines across 18 files), the issue was quickly identified and resolved, reducing potential impact on production.

Summary

Merge multiple DB migrations into single file

Consolidates separate database schema migrations into a single migration, reducing conflicts and simplifying deployment.

Health Assessment

• • Rapid merge after minimal review indicates confidence in changes; large migration consolidation reduces future merge conflicts.

Summary

Bump CLI version to 0.0.50

Update the CLI package version and regenerate the man page to reflect new command groups, ensuring users have up‑to‑date documentation.

Health Assessment

• • Rapid version bump and documentation update merged within 10 hours, indicating minimal review friction and low risk.

Summary

Fix page editor lock ownerId and conflict self‑heal

Ensures reliable lock handling and automatic conflict recovery for page editor saves, preventing intermittent 409 CONFLICT errors and improving user experience. This reduces save failures and improves collaboration stability.

Health Assessment

• • Fast turnaround with minimal rework; AI‑assisted code generation helped resolve complex lock logic quickly.

Summary

Flatten DevDock panel chrome and stop bar overlap

Improves DevDock UI by flattening panel chrome and preventing dock bar from overlapping open panels, enhancing visual consistency. This polish ensures inspectors render cleanly and the dock no longer obscures content.

Health Assessment

• • Rapid review and merge with minimal changes indicates low complexity and high confidence in UI polish.

Summary

Use primary button for update toast

Improves user experience by making the primary update action stand out in the update notification toast, reducing confusion and speeding up updates.

Health Assessment

• • Quick style tweak with minimal code changes, fast review and merge, low risk to production

Summary

Health Assessment

Summary

Align disabled model action with notice

Improves UI consistency by keeping the recovery action next to its notice, preserving width and ellipsis behavior.

Health Assessment

• • Quick style tweak with minimal changes, merged within 2 hours.

Summary

Separate overview from workspace panel

Decouples lightweight conversation overview from the full workspace panel, enabling faster access to files, skills, and documents for users.

Health Assessment

• • Rapid 3.5‑hour cycle with three automated review comments shows high developer velocity, but the large code churn and multiple review rounds raise risk of integration issues.

Summary

Fix bulk topic deletion menu actions

This PR removes bulk topic deletion options from the UI to prevent accidental mass deletions, enhancing data safety for users.

Health Assessment

• • Quick fix with minimal changes and no review friction, indicating a straightforward bug resolution.

Summary

Reuse lobe-ui shiny text style

Replace duplicated CSS animation with compositor-friendly shared style, improving rendering performance and reducing paint overhead.

Health Assessment

• • Merged within 0.9 hours, indicating quick review and minimal friction. Small code changes focused on performance.

Summary

Merge model and reasoning effort into single selector

Simplifies the chat input UI by consolidating model selection and reasoning effort into one menu, improving usability and reducing clutter for users.

Health Assessment

• • The PR was reviewed and merged within 4.2 hours, with a single commit and minimal code churn, indicating a smooth, low-risk change.

Summary

Converge agent-testing into acceptance skill

Unifies duplicate acceptance logic, removes redundancy, and standardizes interaction cost measurement, improving maintainability and consistency. This streamlines skill management and ensures consistent reporting across environments.

Health Assessment

• • The PR had a rapid cycle time and quick review, but required multiple commits to address feedback, indicating moderate iteration intensity.

Summary

Polish acceptance review experience

Enhances user experience for acceptance reviews, streamlining workflow and improving onboarding for users.

Health Assessment

• • Rapid review and merge with minimal iterations indicates high confidence in the changes, though the large scope suggests careful monitoring for downstream impact.

Summary

Fix acceptance checks to reject invalid evidence

Ensures AI acceptance reviews only approve when sufficient evidence is provided, preventing false positives from blank or error screenshots. This improves reliability of the AI review process.

Health Assessment

• • Quick fix with minimal changes and immediate review, indicating low complexity and high confidence.

Summary

Return URLs for created resources

Adds URLs to CLI outputs for resources created, improving traceability and user experience.

Health Assessment

• • Fast turnaround with AI-assisted review and minimal rework, but the PR spans 17 files and 346 lines, indicating a substantial feature addition.

Summary

Add home usage widget slot and refine execution target

Adds a new widget slot for home usage metrics and improves per-user execution target resolution, enhancing user customization and reliability.

Health Assessment

• • PR completed quickly with AI assistance, but involved substantial code changes across multiple locales and backend tests, indicating a significant feature addition with moderate review friction.

Summary

Unify desktop update toasts and text buttons

Standardizes update notifications across renderer and downloaded updates, improving consistency and reducing clutter for non-developer users.

Health Assessment

• • Quick AI-assisted style change with minimal code churn and a single review cycle, indicating a low-risk, high-efficiency update.

Summary

Forbid hard-wrapped prose in markdown evidence

Prevents ragged rendering of agent-authored markdown in chat mode, improving readability for users.

Health Assessment

• • Merged within 36 minutes with no review, indicating a straightforward documentation fix with minimal impact.

Summary

Fix acceptance: refuse round writing to accepted check

Prevents duplicate acceptance of checks, ensuring data integrity and accurate review status.

Health Assessment

• • The PR was merged within 1.4 hours with a single review comment, indicating a straightforward bug fix with minimal rework.

Summary

Add credits slot to chat input control bar

Adds a UI slot for displaying remaining credits next to the approval mode chip, enabling builds to show balance indicators. The open-source default remains unchanged.

Health Assessment

• • Single-commit PR with AI assistance, minimal changes, fast merge.

Summary

Remove failed-check feedback actions

Simplifies the acceptance flow by eliminating unnecessary feedback and actions, improving user experience and reducing code complexity.

Health Assessment

• • Merged within 0.9 hours with minimal changes, indicating a smooth review process and low implementation risk.

Summary

Polish Browser resize and tool surfaces

Improves user experience by eliminating resize artifacts and adding intuitive browser tool surfaces, enhancing productivity and interface consistency.

Health Assessment

• • Rapid iteration with minimal review, indicating high confidence in code quality and clear scope.

Summary

Remove redundant issue marker from comment

Cleaned up code comments to reduce noise, improving maintainability.

Health Assessment

• • Quick review and single commit indicate a straightforward maintenance task; the long cycle time reflects scheduling rather than complexity.

Summary

Fix chat image layout during loading

Ensures consistent image dimensions and layout while images load, improving user experience and preventing layout shifts.

Health Assessment

• • Quick fix with minimal rework, AI review provided suggestions, merged within hours.

Summary

Improve test performance and fix stream tests

Reduces test suite runtime by ~110s and strengthens stream test assertions, improving reliability and developer experience.

Health Assessment

• • The PR made significant performance gains in the test suite and added robust assertions to previously weak stream tests, with minimal review friction and a fast merge cycle.