Pull Request Explorer

Summary

Add ignore action to error briefs

Users can dismiss error briefs without rerunning tasks, improving user experience and reducing unnecessary work.

Health Assessment

• • Quick turnaround with minimal review, indicating a straightforward change with moderate scope.

Summary

Fix server-owned memory embedding runtime

Enables trusted server callers to supply embedding runtime, uses server credentials instead of user key vaults, and adds test coverage.

Health Assessment

Summary

Fix server userMemories to return ingestion results

Return per‑session LoCoMo ingestion results and compute total inserted parts, adding route coverage for multi‑session ingestion output, improving accuracy of memory ingestion reporting for users.

Health Assessment

• • Quick fix with minimal rework; AI review provided fast feedback; PR merged within 0.3 hours, indicating low complexity and high confidence in the change.

Summary

Fix CSP violations by replacing unsafe templates

Removes unsafe new Function usage, ensuring CSP compliance and improving security without affecting user experience.

Health Assessment

• • Quick, single‑round fix with minimal rework; no impact on end‑user functionality.

Summary

Fix agent signal primary skill documents

Ensures both skill bundle and skill index are treated as primary skill documents, improving agent signal handling.

Health Assessment

• • Merged within 20 minutes with minimal changes; AI review bots provided feedback but no code modifications were required.

Summary

Fix missing user id context in tool outcome

Ensures tool outcomes include user context, improving signal accuracy and debugging.

Health Assessment

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

Summary

Fix trailing suffix after partial deltas

Ensures complete message persistence for the Claude Code adapter, preventing truncated content in production.

Health Assessment

• • Fast review and merge with minimal iterations; AI review bots provided quick feedback.

Summary

Add offline_access scope and guard expiresIn default

Fixes silent session expiry by adding offline_access scope and ensuring token expiration defaults, improving user experience by preventing forced re-login.

Health Assessment

• • Quick AI-assisted fix with minimal changes, single review comment, merged within 0.3 hours.

Summary

Release 20260513: Canary Backlog Merge

Ship accumulated canary work into main, adding agent/task improvements, integrations, desktop polish, and reliability fixes.

Health Assessment

• • Rapid release of many changes with minimal review time indicates high confidence but high risk due to large scope.

Summary

Add spark-x2-flash support

Adds support for the Spark X2 Flash model, enabling deploymentName usage and updating base URLs for X series models to improve compatibility with Xfyun’s API.

Health Assessment

• • Fast 3.5‑hour cycle, 0.1‑hour review, minimal comments, small to medium scope, quick merge with low risk.

Summary

Persist agent operations to database table

Adds persistence for agent operations, enabling long‑term tracking beyond Redis TTL, improving reliability and analytics.

Health Assessment

• • Merged in under an hour with a single review comment, indicating a straightforward implementation with minimal friction.

Summary

Add Slack MPIM support and fix Discord DM issue

This PR adds support for Slack MPIM and resolves a Discord DM problem, enhancing the chat integration features.

Health Assessment

• • The PR has a large scope with 20 files changed, which may indicate a complex change.
• • The cycle time is relatively short at 1 hour, suggesting efficient review and merge processes.

Summary

Fix hetero-agent AskUserBridge response events

Ensures AskUserBridge emits events on all terminal paths, preventing UI form stuck and errors.

Health Assessment

• • Quick turnaround with minimal review, indicating a straightforward bug fix; AI-assisted code generation likely contributed to speed.

Summary

Move sub‑agent dispatch from lobe‑gtd to lobe‑agent

Consolidates sub‑agent dispatch logic into the lobe‑agent package, reducing duplication and clarifying responsibilities across the codebase.

Health Assessment

• • Fast 10‑hour cycle and single review round show efficient refactor. The large code movement is contained within a single package, minimizing impact on other components.

Summary

Refactor to use built-in HtmlPreview component

Replaces a custom HTML preview component with the built‑in @lobehub/ui HtmlPreview, cutting code, improving inline preview, and simplifying maintenance.

Health Assessment

• • Fast cycle time (14.7h) and zero review lag indicate smooth integration; AI assistance helped streamline the refactor with minimal rework.

Summary

Translate non-English comments to English

Improves code readability and maintainability for international developers by translating comments to English, ensuring consistency across the codebase.

Health Assessment

• • Quick turnaround with no rework; AI-assisted translation streamlined the process.

Summary

Fix blank user bubble when sending placeholder hint

Resolves issue where user bubble appears blank when sending daily-brief hint from home input

Health Assessment

• • Quick review and merge, low complexity change

Summary

Add agent_operations table

Adds a persistent table for agent runtime operations, enabling long‑term analytics and auditability beyond transient Redis storage.

Health Assessment

• • Large schema change merged quickly after review, but long cycle time indicates extended development period.

Summary

Fix synthetic scroll shrinking spacer

Prevents UI glitches in chat by ensuring spacer remains intact during synthetic scrolls. Improves user experience by maintaining consistent layout.

Health Assessment

• • Quick review but long cycle time suggests scheduling delays; minimal rework indicates a straightforward fix.

Summary

Refactor agent runtime into modular components

Simplifies AgentRuntimeService by extracting core logic into dedicated modules, improving maintainability and testability without changing behavior. This refactor reduces complexity and enhances code clarity.

Health Assessment

• • The PR involved a large refactor of a 2000+ line service, split into three modules, with minimal review friction but a long cycle time due to extensive changes.

Summary

Fix task card date localization

Ensures task card dates render correctly in all UI languages by decoupling date formatting from dayjs global locale, improving internationalization consistency.

Health Assessment

• • Quick fix with minimal changes, AI-assisted review, no major rework.

Summary

Cap response body size to prevent OOM

Prevents serverless function crashes by limiting memory usage during web crawling, ensuring stable operation under heavy load.

Health Assessment

• • The PR was reviewed immediately and resolved in a single commit, indicating a straightforward fix. The change involves a moderate codebase size and addresses a critical memory issue, warranting a medium risk assessment.

Summary

Fix desktop onboarding auth success state

Ensures the desktop onboarding flow correctly displays a success screen after OIDC authorization, improving user experience and reducing confusion.

Health Assessment

• • Quick fix with minimal changes, reviewed by AI tools, merged within a day.

Summary

Add user feedback card and polish task UI

Enhances task drawer UI with user feedback rendering, cleans up task card, and adds Run Now context menu, improving user experience.

Health Assessment

• • Fast cycle time of 1.9 hours, AI-assisted code, minimal review iterations, low risk

Summary

Add missing i18n keys to 16 locales

Adds missing translation strings for task scheduling and review features across 16 locales, ensuring consistent user experience.

Health Assessment

• • Fast review and merge with minimal changes, low risk to production.

Summary

Polish desktop header icons and sidebar density

Improves UI density and consistency by resizing header icons and tightening sidebar spacing, and refactors task menus to use base‑ui components.

Health Assessment

• • Quick turnaround with minimal review friction; UI polish and refactor completed within 2 hours.

Summary

Fix Docs Image

Corrects an image reference in the getting‑started guide, ensuring accurate visual guidance for users.

Health Assessment

• • PR was reviewed and merged within minutes, indicating minimal friction and low risk.

Summary

Add May 11 weekly changelog

Adds a new weekly changelog entry for May 11, documenting the Task feature and linking to the Task guide, improving user onboarding and documentation consistency.

Health Assessment

• • The PR was reviewed immediately and merged within 35 hours, with a few follow-up commits to refine documentation. The use of AI co‑authorship suggests automated drafting of the changelog content.

Summary

Fix Windows CLI agent detection for desktop

Ensures Windows npm-installed CLI agents like Claude and Codex are correctly detected by resolving .cmd shims, improving reliability of CLI agent availability in the desktop app.

Health Assessment

• • Fast review cycle with AI-generated code, minimal friction, quick merge.

Summary

Translate non-English comments to English in CLI

Ensures code comments are in English, improving maintainability and developer onboarding.

Health Assessment

• • Quick review and merge, minimal changes, no functional impact.

Summary

Add hover revert button to discard file changes

Users can now quickly discard per-file changes in the review panel, improving workflow efficiency.

Health Assessment

• • Rapid approval with minimal review, indicating a straightforward UI enhancement.

Summary

Add optimistic create/delete and inline rename

Provides instant UI feedback for document tree operations, improving user experience by reducing perceived latency.

Health Assessment

• • Fast cycle time and minimal rework; AI review offered suggestions but no major changes were required.

Summary

Update Task page placeholder copy

Updates placeholder text on Task page to be more concise and guiding, improving user experience.

Health Assessment

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

Summary

Standardize header action icon sizes

Improves visual consistency of navigation icons, enhancing user experience.

Health Assessment

• • Fast review and merge with minimal changes; no rework or blockers.

Summary

Add dev-only feature flag override panel

Provides developers a UI to toggle feature flags at runtime, streamlining testing and reducing the need for environment changes.

Health Assessment

• • Fast review and merge with minimal iterations; large code addition confined to dev tooling, posing negligible production risk.

Summary

Expose lobe-task and add setTaskSchedule

Users can now enable the lobe-task tool and configure recurring execution schedules, improving automation and control.

Health Assessment

• • Rapid approval with minimal review indicates high confidence in the changes.

Summary

Add service model assignments settings

Users can now configure default agent models and system task models via a new settings page, improving customization and reducing confusion.

Health Assessment

• • Fast cycle time and immediate review indicate minimal friction and high confidence in the change.

Summary

Reset pendingLoginMethod on auth failure/cancel paths

Fixes a state‑management bug that could leave users stuck on the login UI after authentication failures or cancellations, improving user experience.

Health Assessment

• • The PR was reviewed and merged within 13 hours with minimal changes, indicating a smooth and efficient process.

Summary

Refactor SPA dev gating to __DEV__

Switches SPA code to use Vite's __DEV__ define, enabling better dead-code elimination and consistent dev environment handling.

Health Assessment

• • Fast review and merge with minimal changes; no blockers or complex iterations.

Summary

Refactor unified service structure across agent modules

Streamlines agent signal, prompts, database, and tool integration, improving maintainability and consistency.

Health Assessment

• • Fast cycle time and minimal review iterations indicate efficient refactor with low friction.

Summary

Cap image binary to 3.75MB for Anthropic limit

Ensures image uploads stay within Anthropic's 5MB base64 limit by capping binary size, preventing provider errors and improving reliability.

Health Assessment

• • Quick fix with minimal changes, no rework, low risk.

Summary

Fix scheduler, hotkey, comment, TodoList polish

Improves task scheduling reliability, user interaction, and UI clarity, enhancing productivity.

Health Assessment

• • Rapid AI-assisted iteration produced many small commits; review was immediate and minimal, but the high file count indicates significant scope and moderate risk of integration issues.

Summary

Remove stale cron entry from CLI man page

The CLI's generated man page mistakenly listed a removed cron subcommand, causing confusion. Regenerating the man page cleans up the documentation and ensures accurate help output.

Health Assessment

• • Fast turnaround with minimal changes and no review friction.

Summary

Add word wrap toggle to tool arguments display

Enables users to toggle word wrap for long tool argument values, improving readability and reducing truncation.

Health Assessment

• • Fast review and merge with minimal changes, indicating low complexity and high confidence.

Summary

Fix sidebar agent addition

Corrects sidebar behavior when adding agents, improving user experience.

Health Assessment

• • Quick fix with minimal iterations, AI review comment but no changes required.

Summary

Unify ActionIcon sizing and improve TodoList encapsulation

This refactors UI components to ensure consistent icon sizing and better component encapsulation, improving visual consistency and maintainability. No functional changes, purely visual improvements.

Health Assessment

• • Quick visual refactor with minimal review, indicating low complexity and high confidence in changes.

Summary

Polish recommended task templates UX

Enhances user experience for recommended task templates by adding refresh, icon prioritization, inline skill authentication, detailed modal view, and a hero landing on the empty tasks page, improving task creation flow and engagement.

Health Assessment

• • Rapid iteration with minimal review time indicates efficient process, but large code churn suggests significant changes that may increase risk.

Summary

Fix infinite render loop by replacing ScrollShadow

This change eliminates a React error that caused desktop app crashes, improving stability for users with long messages.

Health Assessment

• • The PR was merged within 24 minutes with no review back‑and‑forth, indicating a straightforward, low‑risk fix.

Summary

Fix heteroFinish lifecycle trigger

Ensures task lifecycle transitions correctly when cloud sandbox agent completes, preventing tasks from staying stuck in running state.

Health Assessment

• • Quick AI‑assisted fix resolved a task status hang in the cloud sandbox, with minimal code changes and no review delays.

Summary

Add sandbox persistence and GitHub push rules

Adds documentation to enforce sandbox persistence and GitHub push requirements for cloud context, ensuring code changes are saved and authenticated.

Health Assessment

• • Fast cycle time with minimal rework; documentation update completed in under an hour with only a single comment review.