Pull Request Explorer

Summary

Fix synthetic tool failure diagnostics

Adds detailed reason and tool info to synthetic failure payloads, improving debugging and retry logic.

Health Assessment

• • Quick review and minimal rework indicate a straightforward bug fix with clear test coverage.

Summary

Add project acceptance workspace

Introduces an Acceptance tab within projects, allowing users to view and manage acceptance items directly in the project context, enhancing workflow visibility and reducing onboarding friction.

Health Assessment

• • The PR was reviewed and merged within 31 hours with a single commit and minimal comments, indicating a smooth, low‑friction process.

Summary

Upload PR previews to sibling preview Workers

Enables reliable preview deployments and accurate attribution for Cloudflare Workers, keeping production deployments within rollback limits.

Health Assessment

• • Quick review and merge, minimal rework, small to medium scope, fast cycle time.

Summary

Add MetricModel and TRPC router

Adds a data‑access model and server surface for metrics, enabling idempotent series creation, point tracking, and scoped API access.

Health Assessment

• • Fast cycle time and minimal rework indicate efficient review, but the large code addition suggests significant impact and potential for downstream complexity.

Summary

Fix page editor analytics crash

Removes an analytics event that caused the DocumentModal to crash, ensuring pages open correctly.

Health Assessment

• • Quick, single-file change that resolved a runtime crash with minimal risk.

Summary

Show CI status on PR icons

Adds visual indicators of CI build status to pull request icons in the topic sidebar, improving visibility of build health for users.

Health Assessment

• • Fast review cycle with AI-assisted review enabled a quick merge; minimal rework and low complexity.

Summary

Align loading skeletons with settled layouts

Improves UI consistency by matching loading placeholders to final layouts, enhancing user experience during data fetches.

Health Assessment

• • Quick review and minimal rework indicate a smooth integration of UI improvements.

Summary

Drop OG metadata from auth SPA SEO injection

Removes Open Graph and Twitter Card tags from authentication pages to reduce payload and focus on essential SEO metadata, improving page load and privacy.

Health Assessment

• • Quick refactor with minimal changes and immediate approval, indicating low complexity and high confidence.

Summary

Show token usage when cost unavailable

Provides users with token usage visibility when pricing data is missing, improving transparency and trust.

Health Assessment

• • Quick turnaround with minimal changes and no review comments indicates a straightforward UI tweak.

Summary

Promote evidence alt text to fold row title

Improves clarity of acceptance reports by using custom alt text as row titles, reducing duplicate information.

Health Assessment

• • Fast turnaround with single review and minimal rework suggests straightforward UI tweak.

Summary

Fix workspace directory picker for local device override

Ensures workspace agents with local device overrides display a working directory picker, allowing users to select directories and improving configuration flexibility.

Health Assessment

• • Fast turnaround with minimal code churn, no review required, low risk of regressions.

Summary

Improve device gateway error messaging

Enhances user and model clarity by mapping gateway errors to actionable messages, reducing retry confusion and improving troubleshooting.

Health Assessment

• • Rapid review and merge with minimal iterations indicates high confidence in changes; large code churn but well‑tested.

Summary

Fix task list pagination and virtualization

Ensures all tasks are displayed and improves performance by virtualizing the list, preventing missing tasks and enhancing user experience.

Health Assessment

• • AI-assisted code with rapid review cycles; multiple review rounds but resolved within 4.2 hours, indicating efficient collaboration and moderate risk for a large code change.

Summary

Import SkeletonText from base-ui to fix lint

Fixes CI lint failures by replacing antd Skeleton with base-ui SkeletonText, ensuring consistent styling and enabling successful builds for canary-based PRs.

Health Assessment

• • Quick AI‑assisted change resolved lint issue, enabling CI to pass for canary‑based PRs.

Summary

Refactor Skeleton imports to base-ui

This refactor removes legacy Skeleton imports, preventing lint errors and ensuring consistent UI components across the application.

Health Assessment

• • The PR was completed in under 17 hours with minimal changes, indicating a straightforward, low-risk refactor with no significant review friction.

Summary

Retry network empty completions

Retries first‑party model completions that fail with network errors, reducing service interruptions and improving user experience.

Health Assessment

• • Fast turnaround with minimal rework; PR merged within 4.3 hours, indicating smooth review and low complexity.

Summary

Recenter graph after panel resize

This fix ensures the goal graph remains centered when the UI panel is resized, preventing disorientation and maintaining a consistent user experience.

Health Assessment

• • Fast turnaround with minimal changes and no review needed, indicating low complexity and risk.

Summary

Refactor feature flag evaluation to user ID only

Eliminates email-based lookups in feature flag evaluation, cutting runtime overhead and simplifying the admin UI.

Health Assessment

• • Merged in under 2 minutes with no review comments, indicating smooth approval. The refactor removes email handling from feature flag logic, reducing runtime cost and simplifying the admin UI.

Summary

Fold UI members into vendor core for performance

This PR reorganizes UI module imports to reduce first‑screen bundle size, cutting chunks from 117 to 67 and improving load performance.

Health Assessment

• • Quick turnaround with minimal changes, indicating a straightforward refactor with low risk.

Summary

Replace app shell logo with brand text and update loading hint

Updates the UI to use brand text instead of logo and improves loading hint for better user experience.

Health Assessment

• • Quick turnaround with minimal changes, indicating straightforward UI tweak.

Summary

Fix tool routing to prioritize Knowledge Base

Ensures file lookup requests use the Knowledge Base tool instead of Agent Documents, improving reliability and speed for users.

Health Assessment

• • Quick fix with minimal changes, reviewed promptly, no major rework.

Summary

Rename Work to Task across goal system

This refactor updates terminology from Work to Task throughout the goal management system, improving clarity and consistency without affecting database schema.

Health Assessment

• • Quick turnaround with AI assistance, minimal review friction, and a large refactor across many files.

Summary

Move slug format rules into shared constants

Centralizes slug validation logic to avoid type‑check failures in deployments, improving stability and maintainability. Ensures OSS and custom deployments share consistent validation rules.

Health Assessment

• • PR merged immediately after creation with no review, indicating a straightforward fix.

Summary

Style: Update agent session label in switch-cwd dialog

Updates UI text to use generic 'Agent session' instead of referencing Claude Code, improving clarity across all agents.

Health Assessment

• • Quick style fix with minimal changes, AI-assisted, merged within 14 hours.

Summary

Resolve internal entity links into titles and surface underline

Improves findability of internal links in chat by rendering entity titles instead of raw URLs and making underlines visible, enhancing user experience.

Health Assessment

• • Quick review and merge with minimal changes, indicating a straightforward implementation that was well understood by reviewers.

Summary

Refactor AgentTasks: Remove run summary/output labels

Simplifies UI by removing redundant captions, improving clarity and reducing visual noise.

Health Assessment

• • Quick refactor with minimal changes, merged within 2.3 hours, indicating smooth review and low friction.

Summary

Add node-kind filter to exploration map legend

Enables users to toggle visibility of node kinds in the goal graph, improving focus and clarity.

Health Assessment

• • Quick review and merge with minimal rework, indicating a straightforward UI enhancement.

Summary

Add agent sharing via link for visitors

Enables creators to share agents with visitors through a link, allowing direct chat without marketplace publishing, improving user engagement and reducing onboarding friction.

Health Assessment

• • The PR required extensive back‑and‑forth revisions, with multiple AI‑generated review comments and a high number of commits after the first review, indicating significant complexity and potential risk.

Summary

Improve upload performance by offloading hashing

This PR moves SHA‑256 hashing to background workers, eliminating UI freezes and speeding up large file uploads.

Health Assessment

• • Rapid review and merge with minimal iterations indicate high confidence in the performance improvements and low risk to the codebase.

Summary

Hide scope toggle from running tasks

Removes the scope toggle from the running‑tasks card to declutter the workspace home, improving user focus.

Health Assessment

• • The PR was reviewed and merged within 12 hours, with a single AI‑generated review comment and no further changes, indicating a smooth, low‑risk process.

Summary

Fix no-active-device failures for agents

Ensures agents can recover from missing device errors, preventing stalls and improving reliability for users interacting with device-capable tools.

Health Assessment

• • Merged within 3 hours with a single commit and no review comments, indicating a straightforward, low-risk bug fix.

Summary

Add workspace commands and persisted scope

Introduces a full set of CLI commands for workspace management, allowing users to list, view, switch, and create workspaces, and persists the selected workspace scope for subsequent operations.

Health Assessment

• • The PR was reviewed and merged within 34.5 hours with a single review comment, indicating a smooth review process despite the large code addition.

Summary

Add fullscreen map drill-down and resolve problem nodes

Enables users to drill down into goal maps from fullscreen view and ensures achieved goals no longer show unresolved problem nodes, improving clarity and user experience.

Health Assessment

• • Quick turnaround with minimal review, indicating straightforward implementation and high confidence in changes.

Summary

Fix reading large PDFs

Enables reading PDFs larger than the default 10 MiB guard by adding a 50 MiB cap before loading into memory, with regression tests to ensure stability.

Health Assessment

• • Fast cycle time of 1.3 h and an AI‑assisted review led to a smooth, low‑friction merge.

Summary

Batch move multi-selection into a project

Enables bulk moving of selected acceptance items into a project, streamlining workflow and reducing repetitive actions for users.

Health Assessment

• • Fast review turnaround and single commit indicate smooth integration, but the large code change size suggests moderate risk for downstream impact.

Summary

Fix desktop-only electron SWR hooks on web

Prevents web crashes by gating electron-specific hooks, improving stability and user experience.

Health Assessment

• • Rapid resolution with a single commit and minimal review, indicating low complexity and risk.

Summary

Fix re‑dispatch of delivered Work during verification

Prevents duplicate task retries and data loss, ensuring accurate goal completion and avoiding unnecessary charges.

Health Assessment

• • Fast turnaround with minimal rework; AI-assisted code generation streamlined the fix.

Summary

Add generic metrics time-series tables

Introduces database tables for generic metrics, enabling future charting and analytics across multiple subjects.

Health Assessment

• • Merged in under 9 hours with a single commit, indicating quick review, but the large schema change carries high risk for downstream systems.

Summary

Add cloud sandbox to auto mode

Enables the cloud sandbox tool in auto mode, improving credential reachability and user experience for agents.

Health Assessment

• • Rapid AI-assisted review and multiple post-review commits indicate thorough refinement, but the large scope and several iterations suggest moderate risk.

Summary

Fix CLI detection cache reuse on launch

Improves performance by reusing cached CLI detection, reducing repeated shell startups and preventing timeouts during message processing.

Health Assessment

• • Quick fix with minimal rework, AI-assisted review accelerated approval.

Summary

Fix React artifact preview for large apps

Ensures Sandpack preview remains functional for artifacts exceeding 16 KiB by switching to Babel-based transforms, preventing esbuild protocol desyncs and improving reliability.

Health Assessment

• • Quick fix with minimal changes; AI-assisted review helped identify and resolve the issue rapidly.

Summary

Add structured acceptance criteria and goal chat

Users can now define structured acceptance criteria, chat with goal context, and receive accurate liveness updates, improving clarity and reliability.

Health Assessment

• • The PR involved extensive code changes across backend and frontend, with six review rounds and multiple iterations, yet it was merged within 8 hours, indicating a focused and efficient review process.

Summary

Fix Vite barrel import namespace handling

Ensures Vite builds correctly by forwarding empty barrel imports as namespaces, preventing parse errors in production. This resolves build failures in cloud deployments caused by empty imports.

Health Assessment

• • Quick resolution with minimal changes; likely a straightforward fix with no major risk.

Summary

Fix tree moves and root sidebar creation

Resolves drag‑and‑drop and sidebar creation bugs, ensuring correct tree synchronization and root‑level resource creation, improving user experience and data consistency.

Health Assessment

• • Quick turnaround with minimal review friction; the PR addressed critical UI bugs with extensive changes but was resolved efficiently.

Summary

Improve DevDock compact performance metrics

Enhances visibility of performance metrics in constrained UI, reducing alert fatigue and improving developer experience.

Health Assessment

• • Fast cycle time and minimal rework indicate a well‑executed feature addition with low risk.

Summary

Add Gemini 3.8 Flash and fix React preview

Adds Gemini 3.8 Flash defaults and locale descriptions, enabling accurate model behavior. Fixes React artifact preview failures in Sandpack, improving developer experience.

Health Assessment

• • Fast turnaround with minimal rework; quick review and merge indicates low complexity.

Summary

Fix CI size baseline selection

Ensures CI gates use correct baseline artifacts, preventing false positives in bundle size checks.

Health Assessment

• • Quick fix with minimal rework, fast review, low risk.

Summary

refactor(skeleton): migrate to base-ui Skeleton and unify loading layer

Streamlines the UI loading experience by consolidating skeleton components into a single, reusable base-ui API, eliminating duplicate loading states and improving performance and maintainability.

Health Assessment

• • The PR was reviewed quickly (0.1h) but required many iterations and automated Codex review comments, indicating significant rework to satisfy quality checks. The moderate code churn (≈480 lines) and 20 files suggest a substantial refactor with potential impact on downstream components.

Summary

Fix OIDC redirect to use relative origin

Ensures OIDC consent and callback redirects use relative URLs so the browser stays on the correct origin, preventing SessionNotFound errors when using lobehub.com domain.

Health Assessment

• • Quick fix with minimal changes, no major rework, resolved redirect issue in a single commit.

Summary

Add Qwen3.8 Max 0902 description to locales

Adds a description key for the new Qwen3.8 Max 0902 model across all supported locales, enabling consistent user-facing information.

Health Assessment

• • Fast review and merge with minimal changes, low risk