Pull Request Explorer

Summary

Stream live sub-agent progress to parent

Enables real‑time progress updates for sub‑agent operations, improving visibility and user experience during long‑running tasks.

Health Assessment

• • Fast turnaround with minimal review iterations indicates high confidence in the change.

Summary

Fix hourly extraction task tracking

Adds batch-level async task tracking and cancellation support for memory extraction, improving reliability and reducing failures for users.

Health Assessment

• • Large-scale bugfix with extensive test coverage and automated review, but a 164‑hour cycle time indicates bottlenecks in the review or approval process.

Summary

Open message link in side browser on hover

Adds a hover‑activated action that opens chat message links in the in‑app side browser, enhancing user experience by keeping browsing within the app.

Health Assessment

• • Fast review and single iteration indicate a low‑risk, straightforward enhancement.

Summary

Fix Claude Code error reporting

Improves error visibility for users running Claude Code, enabling quicker troubleshooting and reducing support overhead.

Health Assessment

• • Fast turnaround with minimal changes, quickly addressing a user-facing issue and enhancing reliability.

Summary

Stabilize virtualized chat scroll assertions

Fixes e2e test failures by correcting scroll assertion logic, ensuring reliable UI behavior. No product code changes, only test improvements.

Health Assessment

• • Rapid fix with minimal rework; test-only change improved stability and CI reliability.

Summary

Fix desktop internal link handling

Ensures internal links open within the app portal instead of the system browser, improving user experience and keeping users in the application context.

Health Assessment

• • The PR was merged in under an hour with a single commit and no review rounds, indicating a straightforward bug fix with minimal friction.

Summary

Add context quality linter and golden payload tests

Introduces deterministic linting of LLM payloads and regression tests to ensure payload consistency, enhancing agent reliability and reducing error rates.

Health Assessment

• • The PR was reviewed and merged within 10 hours, with a single commit and minimal back‑and‑forth, indicating a smooth and efficient process.

Summary

Refactor agent runtime for client tool execution

Streamlines agent tool execution by consolidating client handling, reducing duplication and improving reliability.

Health Assessment

• • Fast review cycle and single commit after review indicate a smooth, low-risk refactor.

Summary

Fix gateway sync issue

Corrects synchronization errors in the gateway service, improving reliability of message routing.

Health Assessment

• • The PR was reviewed and merged within an hour, with minimal rework and a single AI‑powered review comment, indicating a smooth, low‑risk fix.

Summary

Fix workspace topic sharing functionality

Corrects broken sharing for workspace‑scoped topics, ensuring proper message loading, permission checks, and audit logging.

Health Assessment

• • Fast cycle time and minimal rework; automated AI review helped identify edge cases early, leading to a smooth merge.

Summary

Add 404 fallback for private agents/files

Provides a user‑friendly error page when a previously public agent, group, or file is made private or deleted, preventing broken links and improving user experience.

Health Assessment

• • The PR was merged quickly (8 h) but involved many commits and review comments, indicating intensive rework on a large codebase.

Summary

Fix creator-only visibility demotion and sidebar reveal

Ensures only creators can demote visibility, preventing data appropriation, and improves UI feedback by revealing sidebar sections after visibility changes.

Health Assessment

• • Fast cycle time (8.3h) and minimal review indicate a straightforward bug fix; AI assistance helped streamline code changes and testing.

Summary

Migrate @lobehub/ui components to base-ui

Refactors the remaining deprecated @lobehub/ui component imports to @lobehub/ui/base-ui, reducing lint debt and improving code consistency.

Health Assessment

• • The PR has a large scope with many file changes, but the review process was relatively smooth with no major issues reported.

Summary

Enable Claude Code to control in‑app browser

Adds MCP tools that let Claude Code drive the desktop in‑app browser, enhancing agent interaction and user experience.

Health Assessment

• • Fast review and merge cycle, but multiple commits after the first review indicate iterative refinement; the large scope across backend and frontend suggests significant changes that warrant careful testing.

Summary

Update failed topic status icon to alert triangle

Adjusts UI icon for failed topics to better reflect alert semantics, improving user clarity.

Health Assessment

• • Quick style tweak merged within minutes, minimal risk.

Summary

Unify share routes behind ShareShell business slot

Consolidates share page and topic routes into a single shell, improving consistency and reducing duplication across the application.

Health Assessment

• • The PR was merged within 4 hours with no review comments, indicating a smooth, low-risk refactor of shared routing logic.

Summary

Add client runtime host adapters for agent

Enables client‑side agent execution with an optimistic chat store, improving responsiveness and reducing server load.

Health Assessment

• • Quick turnaround with minimal rework, indicating a straightforward refactor with clear scope.

Summary

Add comparison evidence support in verify reports

Enables side‑by‑side comparison of verification evidence, improving clarity for users and stakeholders.

Health Assessment

• • Quick review and merge indicate a straightforward change with minimal friction.

Summary

Add archive topics for merged PRs

Enables users to archive topics linked to merged pull requests, keeping topic lists clean and synchronized across agent and group views.

Health Assessment

• • Fast review and minimal rework indicate a smooth, low‑risk process.

Summary

Fix heterogeneous subagent message persistence

Ensures subagent messages are persisted correctly, improving reliability of agent workflows.

Health Assessment

• • The PR was reviewed quickly but required several commits to address feedback, indicating moderate iteration intensity.

Summary

Add link opening in report portal

Enables users to open verification links directly from the report portal, improving workflow efficiency.

Health Assessment

• • The PR was reviewed and merged within 3.2 hours, with an AI-assisted review comment and minimal rework, indicating a smooth and low-risk change.

Summary

Fix cross-mode key pollution in topic groups

Corrects topic group expand/collapse behavior across modes, preventing stale state and improving user experience.

Health Assessment

• • Quick 6.6‑hour turnaround with no review iterations; minimal code churn and no blockers.

Summary

Add agent-driven browser control for desktop

Enables agents to control an in-app browser, supporting both local and cloud execution, improving automation and cross-device workflows.

Health Assessment

• • The PR required multiple iterations after the initial review but was merged within a day, indicating efficient communication and rapid resolution of issues.

Summary

Add resumable hetero continuation scheduling

Enables users to schedule AI topic continuations after usage limits, reducing interruptions and improving engagement.

Health Assessment

• • Fast cycle time and AI-assisted review kept friction low, but the large code churn and many commits suggest a complex change that may carry higher risk.

Summary

Fix ControlBar git status and add Fable 5 model

Ensures worktree topics display git status in the ControlBar and adds the Fable 5 model to the Claude Code selector, improving user experience and expanding model options.

Health Assessment

• • Quick review and single commit indicate a smooth, low‑friction process.

Summary

Refactor: Remove Legacy LLM Turn Executor

Simplifies LLM execution flow by removing legacy transport API and consolidating retry and tracing logic into a single transport layer, improving maintainability and reducing potential bugs. This refactor reduces code churn and streamlines future feature development.

Health Assessment

• • The PR was merged within 2.6 hours with a single commit and no review comments, indicating a straightforward refactor with low perceived risk.

Summary

Fix agent branch refresh after CLI checkout

Ensures the agent’s branch cache stays in sync with CLI operations, preventing stale state and improving reliability.

Health Assessment

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

Summary

Add locales for insufficient credits and tidy error actions

Enhances user experience by providing localized messages for insufficient credits and simplifying error message actions, improving clarity and reducing confusion.

Health Assessment

• • Fast cycle time and minimal review rounds indicate a smooth process; AI review provided concise feedback, leading to quick resolution.

Summary

Fix e2e SSE streaming in LLM mock

This PR corrects malformed SSE payloads in the LLM mock, ensuring tests accurately simulate streaming behavior. It also adds deterministic regression tests and improves test reliability for scroll and conversation features.

Health Assessment

• • Fast review and merge with minimal rework, quickly resolving a latent test‑infrastructure defect.

Summary

Style connector tool names with theme code font

Ensures connector tool names display in a monospace font on both web and desktop, improving visual consistency and readability.

Health Assessment

• • Quick style fix with a single file change, minimal risk and no tests required.

Summary

Stabilize skill policy menu actions

Fixes click reliability for MCP policy actions by coordinating popovers and preventing unintended closures.

Health Assessment

• • Quick fix with minimal changes, merged within 1.1 hours, indicating low complexity and high confidence.

Summary

Update skill policy menu to use action verbs

Improves clarity of skill activation menu, making actions more intuitive for users and reducing confusion between state and action.

Health Assessment

• • Quick, single-commit PR with minimal changes, merged within 2 hours, indicating low complexity and high confidence in correctness.

Summary

Inject workspace scope for device system queries

Ensures workspace-bound agents correctly resolve device system variables, improving reliability of agent prompts. This fixes failures in workspace sessions where system info was missing.

Health Assessment

• • Merged within 2 hours with minimal changes, indicating a straightforward fix and low risk.

Summary

Keep profile breadcrumb left-aligned

Ensures the Agent Profile breadcrumb remains left-aligned regardless of panel state, improving UI consistency and user experience.

Health Assessment

• • Fast, single-commit fix with regression test, minimal impact.

Summary

Restore activated tools across operations

Ensures tools activated in one operation persist into subsequent operations, improving agent continuity and reducing redundant activations.

Health Assessment

• • Fast review and merge cycle with minimal rework; the PR addressed a subtle bug with clear test coverage and no major architectural changes.

Summary

Fix legacy MCP migration and connector config panel

This PR resolves issues with legacy custom MCP migration, preventing duplicate connectors and ensuring proper configuration flow, improving stability for users importing connectors.

Health Assessment

• • Fast cycle time (1h) and minimal rework indicate a smooth review process; the PR addresses a critical bug with moderate code changes and no major risk to production.

Summary

Refactor message usage storage to top-level column

Simplifies message storage by eliminating duplicate metadata, improving performance and maintainability for future feature development.

Health Assessment

• • PR completed in 22 hours with a single commit and no review iterations, indicating a straightforward refactor with minimal risk.

Summary

Refactor LLM finalization into agent runtime

Consolidates LLM finalization logic into the agent runtime, reducing server complexity and improving maintainability for more reliable chat interactions.

Health Assessment

• • Merged within 30 minutes with no review rounds, indicating high confidence in the refactor and minimal risk to existing functionality.

Summary

Gate in-app Browser tab behind Labs toggle

Enables users to control access to the in‑app browser, allowing the feature to mature before exposure and reducing potential confusion or misuse.

Health Assessment

• • Quick single‑commit PR with AI assistance, minimal review needed, fast merge.

Summary

refactor: move LLM retry orchestration into agent runtime

Reorganizes LLM retry logic into runtime, improving maintainability and traceability of AI interactions. This refactor enhances reliability and simplifies future feature extensions.

Health Assessment

• • Merged within 10 hours with no review comments, indicating a straightforward refactor with low complexity.

Summary

Fix add member modal footer spacing

Corrects spacing in the add member modal footer, improving the visual consistency and user experience.

Health Assessment

• • The PR was reviewed immediately and merged after a short cycle, indicating a straightforward UI fix with minimal risk.

Summary

Remove Chinese fallback text from i18n components

Eliminates hardcoded Chinese strings, ensuring consistent localization and preserving UI behavior.

Health Assessment

• • Quick fix with minimal changes, merged within a day.

Summary

Add visibility field to device test factory

Adds missing visibility field to device test factory, fixing type errors that caused release pipeline failures and ensuring reliable builds. This prevents build failures in the Desktop Canary release, improving CI reliability.

Health Assessment

• • PR resolved a type-check issue that broke CI, with minimal changes and no review needed.

Summary

Rename 'completed' to 'archive' in topic UI

Aligns terminology across the topic interface, ensuring consistent status handling and improving user experience.

Health Assessment

• • Rapid review and merge with minimal changes, indicating low complexity and risk.

Summary

Fix Zod 4 type inference for database schemas

Corrects type mismatches caused by the Zod 4 upgrade, ensuring runtime validation and TypeScript type safety for database insert/update operations.

Health Assessment

• • The PR was reviewed and merged within 0.7 hours, with minimal rework and no significant scope, indicating a smooth, low-risk fix.

Summary

Fix prompt save feedback and header layouts

Improves user experience by providing accurate autosave feedback and consistent layout alignment, reducing confusion and errors.

Health Assessment

• • The PR was merged within 0.8 hours with minimal review and rework, indicating a straightforward bug fix with strong test coverage.

Summary

Add workspace device visibility, sharing, and per-user agent preference

Enables fine‑grained device visibility and sharing within workspaces, allowing users to control which devices are visible, share devices across workspaces, and set per‑user agent device preferences, improving collaboration and security.

Health Assessment

• • The PR required many review comments and rework, indicating complex backend logic changes, but was merged quickly, suggesting efficient resolution.

Summary

Fix bot silent issue by recovering empty replies and notifying thread members

This PR resolves a bug where bot replies were missing in IM channels by recovering empty completion replies from the database and ensuring thread members are notified, improving reliability and user experience.

Health Assessment

• • The PR was reviewed and merged quickly, with minimal back‑and‑forth, indicating a smooth process. AI assistance was used to generate the code, but the changes were straightforward and well‑tested.

Summary

Add sortRouterOptions hook to RouterRuntime

Enables hosts to reorder fallback options before each request, improving routing flexibility and resilience.

Health Assessment

• • The PR was reviewed and merged within 1.7 hours with minimal back‑and‑forth, indicating a smooth, low‑risk change.

Summary

Compact heterogeneous model selector UI

Reorganizes the agent selector into concise rows, reducing menu height and improving user navigation.

Health Assessment

• • The PR was merged within 1.5 hours with no review comments, indicating a straightforward UI enhancement with minimal risk.