Pull Request Explorer

Summary

Refresh home recents and inline task status

Enables real‑time updates on the home page by polling recents every 10 seconds and embedding task status directly, reducing API calls and improving user experience.

Health Assessment

• • Merged within 2 hours with no rework; AI review bots provided suggestions but did not affect code quality.

Summary

Fix memory search sanitization for backticks

Corrects a bug that caused memory search failures when backticks were present, ensuring reliable search functionality for user memories.

Health Assessment

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

Summary

Fix user memory errors

Explicitly inject user memory errors to improve error handling.

Health Assessment

• • Quick review and merge process, indicating a straightforward fix.

Summary

Refactor: Merge Agent Marketplace into Web Onboarding

Consolidates onboarding tooling, reduces package sprawl, and adds locale‑aware fetching for a smoother user experience.

Health Assessment

• • Fast review cycle (0h) and minimal rework indicate a clean, low‑risk refactor. The 20‑file, 260‑line change is moderate in scope but well‑contained within the frontend stack.

Summary

Fix TypeScript compilation error

Fixes a TypeScript compilation error in the Discord OAuth service, ensuring the backend compiles correctly and prevents runtime failures.

Health Assessment

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

Summary

Add task template tracking and analytics

Adds client‑side analytics for task template recommendations, enabling better funnel insights and user engagement tracking.

Health Assessment

• • Rapid iteration with minimal review time; large code changes but quick approval.

Summary

Add support for messager integration

Enables users to connect via messenger platforms, expanding communication channels.

Health Assessment

• • The PR was reviewed immediately and merged within 2.7 days, but the large number of commits and file changes suggests significant iteration, though no code changes reduce technical risk.

Summary

Fix tool call timer reset on collapse and expand

Ensures tool call timers persist correctly when the UI collapses or expands, improving reliability of tool execution metrics for users.

Health Assessment

• • Rapid review and merge with minimal comments indicate low complexity and high confidence in the change.

Summary

Fix Review Panel Gating to Effective Working Directory

The review panel now aligns with the agent's effective working directory, ensuring accurate context for review actions. This reduces errors and improves user experience when working across multiple repositories.

Health Assessment

• • Quick turnaround with minimal review comments indicates a straightforward bug fix.

Summary

Fix sanitization of provider tool names

Ensures tool names are sanitized to avoid provider rejections, improving compatibility with strict APIs.

Health Assessment

• • Quick fix with minimal rework; automated review comments were rate‑limit related but did not block the merge.

Summary

Fix task template recommendations visibility

Always show task template recommendations, even when users have multiple unresolved briefs.

Health Assessment

• • Quick merge with no significant review discussion.

Summary

Fix Claude thinking history signature sanitization

Ensures that historical assistant messages are properly sanitized before being sent to Claude, preventing signature mismatches and improving request integrity.

Health Assessment

• • Quick fix with minimal changes, no blockers, fast review cycle.

Summary

Add messenger tables for IM bot integration

Adds database tables and models to support shared‑bot integration across Slack, Telegram, Discord, Feishu, and MS Teams, enabling secure credential storage and user mapping.

Health Assessment

• • Rapid merge with minimal review; large database schema changes may impact deployments and require careful migration testing.

Summary

Simplify onboarding agent identity card

Improves user experience by removing redundant UI elements and clarifying the agent’s intent during onboarding, reducing cognitive load.

Health Assessment

• • Quick turnaround with minimal review, indicating a straightforward UI refactor that required no significant rework.

Summary

Fix gateway input loading during execAgentTask round-trip

Ensures the send button remains in loading state throughout the agent task execution, preventing flicker and maintaining a consistent UI during network round‑trips.

Health Assessment

• • The PR was merged quickly with minimal review, indicating confidence in the changes, but the large code churn across backend and frontend suggests significant refactoring.

Summary

Fix assistant content override rendering

Ensures each assistant message segment displays its intended content, improving accuracy of displayed information for users.

Health Assessment

• • Resolved quickly with a single commit and no review back‑and‑forth, indicating a straightforward fix.

Summary

Integrate BM25 search and docs_* read for inline documents

Adds BM25 search capability and inline document reading to the knowledge base tool, improving search accuracy and enabling direct access to custom documents.

Health Assessment

• • Fast 2.7‑hour cycle with minimal rework; AI review helped surface errors early and keep the PR concise.

Summary

Move tasks to agent sidebar list

Replaces inline task list with a dedicated sidebar section, improving task visibility and navigation for users.

Health Assessment

• • The PR was reviewed and merged within 36 minutes, indicating a smooth, low-friction process with minimal rework.

Summary

Fix onFinish never called during SSE streaming

Prevents chat completions from stalling when the browser tab is backgrounded during SSE streaming by replacing await-on-animation with synchronous flushQueue()

Health Assessment

• • The PR was reviewed and merged quickly, indicating low review friction and a well-understood fix.

Summary

Refactor agent API routes to Hono

Consolidates agent API endpoints into a single Hono application, preserving behavior while improving maintainability and performance.

Health Assessment

• • Merged within 13 hours with minimal review, indicating smooth integration despite large code movement.

Summary

Add lobeAgents markdown tag for inline agent card

Enables LLMs to embed interactive agent cards directly in chat, improving user engagement and reducing navigation friction.

Health Assessment

• • Fast cycle with minimal rework; AI assistance accelerated development.

Summary

Fix sanitize assistant media in Responses input

Prevents invalid assistant media from being sent to providers, ensuring compatibility and reducing errors.

Health Assessment

• • Fast cycle time and immediate review indicate a straightforward bug fix with minimal friction.

Summary

Add clickable agent card after creation

Enables users to quickly access newly created agent sessions via a clickable card, improving workflow efficiency.

Health Assessment

• • Fast turnaround with minimal review; AI assistance streamlined development.

Summary

Fix desktop notification settings display

Adds desktop support for notification settings via cloud webview and regression test, improving user experience.

Health Assessment

• • Quick fix with minimal changes, AI review helped catch issues, merged within minutes.

Summary

Add RecommendTaskTemplates UI and default noop router

Introduces task recommendation UI and a default router for desktop and self‑hosted users, enabling the "Try these tasks" feature in the open‑source build.

Health Assessment

• • Merged within 1.1 hours with no review rounds, indicating a straightforward, low‑risk change despite a large code addition.

Summary

Fix runtime config to hide LobeHub toggle

This change ensures the LobeHub provider toggle visibility respects the backend configuration, preventing users from disabling the provider when it is managed by the platform.

Health Assessment

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

Summary

Fix nullable parsing in memory-user-memory

Corrects activity parsing to allow nullable values, improving data integrity.

Health Assessment

• • Quick fix with minimal changes, reviewed instantly by AI engine, merged within minutes.

Summary

Update brief template style

Enhances UI consistency by updating brief template styling and i18n translations, improving user experience.

Health Assessment

• • Rapid review and merge with minimal iterations; large translation update but no functional impact.

Summary

Health Assessment

Summary

Revert e2e topic group expansion workaround

Reverts a test workaround to restore original topic group collapse behavior, improving test reliability.

Health Assessment

• • Fast merge with minimal changes, no review required.

Summary

Polish onboarding UI and add web‑onboarding tool

Enhances the onboarding experience with a cleaner UI and new web‑onboarding tool components, improving user engagement and reducing friction during initial setup.

Health Assessment

• • Rapid iteration with minimal review; large code changes focused on UI polish and new tool components, indicating a high‑impact but potentially risky update.

Summary

Add inactivity watchdog finalize endpoint and migrate agent-hono

Adds a finalize-abandoned endpoint to clean up orphaned agent runs, ensuring errors are reported and resources are freed, while migrating key agent APIs to the Hono framework for better performance and maintainability.

Health Assessment

• • The PR introduced a large number of changes across many files, but was reviewed and merged quickly, indicating efficient collaboration. However, the extensive scope and multiple commits suggest significant rework, raising potential risk.

Summary

Reduce streaming re-renders via reference stabilization

Improves rendering performance for assistant messages during streaming, cutting unnecessary re-renders and boosting UI responsiveness.

Health Assessment

• • Fast review and merge with minimal rework, but the change touches over a thousand lines across many files, indicating a substantial performance optimization effort.

Summary

Fix editor runtime mutation handlers

Adds mutation handlers to ensure editor state synchronizes correctly after saves, improving reliability and user experience.

Health Assessment

• • The PR was merged quickly after an AI‑assisted review, but required several commits to address test failures and refine synchronization logic.

Summary

Fix preserve nested router runtime id

Preserves inherited runtime id for nested router chains, ensuring provider-scoped pricing lookup.

Health Assessment

• • Quick review and merge with minimal changes.
• • Test coverage was added, ensuring the fix is verified.

Summary

Add grok-4.3 to LobeHub-hosted xAI models

Adds a new AI model card with tiered pricing and disables older generation, enabling users to access the latest model on LobeHub Cloud.

Health Assessment

• • Quick approval with automated review, minimal changes, fast merge.

Summary

Fix xAI Responses API tool stability

Improves reliability of xAI integration by routing chat requests to the Responses API, mapping response formats, sanitizing tool schemas, and updating tests.

Health Assessment

• • Rapid resolution with minimal rework; AI review helped catch issues early.

Summary

Update ChatInput placeholder text for heterogeneous agents

Improves user experience by making placeholder prompts more natural and consistent with Claude Code style.

Health Assessment

• • Quick style update with minimal changes, reviewed and merged within 12 hours, indicating efficient process.

Summary

Persist home sidebar collapse state

Ensures the user’s sidebar collapse state is retained across page reloads, improving the continuity of the user experience.

Health Assessment

• • The PR was reviewed and merged within minutes, indicating a straightforward change with minimal risk and no significant rework.

Summary

Expand visible topic groups in E2E steps

Improve conversation management E2E tests by expanding visible topic groups and handling topic visibility uniformly.

Health Assessment

• • The PR was reviewed and merged quickly, indicating a smooth review process.
• • The small scope and low number of files changed suggest a low-risk change.

Summary

Fix streaming executor agent spy tests

Fixes failing tests for the streaming executor agent spy, improving test reliability and confidence in streaming features.

Health Assessment

• • Quick fix with no review needed, minimal scope, low risk.

Summary

Fix template variable leakage in execAgent context

Ensures server‑side agent context correctly resolves template variables, preventing sensitive data leakage and improving LLM accuracy.

Health Assessment

• • Quick, single‑commit fix with no rework, minimal scope and low risk to production.

Summary

Add app tray visibility setting

Users can now control whether the app tray icon is shown, improving UI customization and reducing clutter.

Health Assessment

• • Fast merge with minimal review, indicating a straightforward change with low risk.

Summary

Fix context compression threshold config

Ensures chat agents use correct compression thresholds, improving performance and reliability.

Health Assessment

• • Fast review and merge with minimal changes indicates low complexity and high confidence in the fix.

Summary

Fix task breadcrumb title truncation

Adjust breadcrumb layout to prevent long titles from being truncated incorrectly, improving UI clarity for task details.

Health Assessment

• • Quick review and merge within minutes, minimal rework.

Summary

Inject user language into task summary chains

Ensures task titles and briefs are rendered in the user’s preferred language, improving localization consistency.

Health Assessment

• • Quick, single‑commit fix with AI review, minimal scope and fast turnaround.

Summary

Fix agent signal skill targeting and auto refresh

Corrects agent signal handling for skill documents and ensures automatic refresh, improving reliability of skill execution.

Health Assessment

• • Rapid review and merge with minimal changes indicates high confidence in the fix.

Summary

Add step-image-edit-2 support

Adds support for the step-image-edit-2 model, enabling new image editing capabilities and exposing additional parameters for image generation.

Health Assessment

• • The PR had a long cycle time but minimal code changes, indicating a straightforward feature addition with quick review.

Summary

Fix bot upload visual refs preservation

Preserves visual media refs for bot-uploaded images and videos, ensuring analyzeVisualMedia can resolve the same refs shown to the model.

Health Assessment

• • Fast cycle time and immediate review suggest a straightforward fix.

Summary

Add May 4 weekly changelog

Adds updated changelog for May 4, documenting new features and improvements for users.

Health Assessment

• • Quick merge after a single review, minimal changes, no code modifications.