Pull Request Explorer

Summary

Add bot attachment support across platforms

Enables bots to send images, videos, and files on Discord, Telegram, Slack, Feishu, LINE, and QQ, expanding cross‑platform messaging capabilities.

Health Assessment

• • Rapid AI‑assisted review and merge within 7.5 hours, but the large code change (≈4,800 lines) introduces high risk of regressions.

Summary

Prefetch agent marketplace templates during onboarding

This change preloads agent marketplace templates during onboarding, reducing wait times and improving user experience.

Health Assessment

• • Rapid review and merge with minimal changes indicates high confidence and low risk.

Summary

Unify TabBar registration into cross-platform route-meta

Centralizes tab metadata into route definitions, simplifying desktop navigation and reducing code duplication.

Health Assessment

• • Rapid review and merge with minimal iterations indicates high confidence in the refactor and low risk to stability.

Summary

Fix bad request for malformed auth JSON

Ensures the authentication API returns a clear 400 error for malformed JSON, improving client reliability and reducing server error noise.

Health Assessment

• • Quick fix with minimal review, indicating straightforward bug resolution.

Summary

Skip Pro Settings step when Klavis disabled

Ensures onboarding flow skips Pro Settings step when server config disables Klavis, improving user experience and reducing unnecessary steps.

Health Assessment

• • PR merged quickly with minimal review, indicating a straightforward bug fix with clear tests.

Summary

Restore mobile padding on Classic onboarding steps

Improves mobile UX by adding horizontal breathing room to Classic onboarding steps, ensuring content is not flush with viewport edges.

Health Assessment

• • The PR was merged within 42 minutes, with no review iterations, indicating a straightforward, low-risk change.

Summary

Add LaTeX extensions to text file types

Ensures LaTeX files are treated as plain text, preventing binary rejection in lobe-local-system.

Health Assessment

• • Quick fix with immediate review and merge, minimal scope and risk.

Summary

Health Assessment

Summary

Adapt agent onboarding UI for mobile

Improves mobile onboarding by adding a dedicated mobile greeting, adjusting layout for small screens, and ensuring the input area remains visible on first paint.

Health Assessment

• • Fast 3.9‑hour cycle with a single commit and no rework; minimal review friction and a straightforward UI enhancement.

Summary

Track create agent modal source analytics

Adds telemetry‑gated analytics for agent creation, enabling better insight into user behavior while respecting privacy settings.

Health Assessment

• • Merged within 45 minutes with no review rounds, indicating a straightforward change; high line count but focused on analytics, minimal risk.

Summary

Fix redundant visual tool calls

Discourages unnecessary visual analysis tool calls by adding a fallback guidance to the Lobe Agent tool system role.

Health Assessment

• • The PR was reviewed and merged quickly, indicating a smooth development process.

Summary

Fix chat input draft restore bug

This fix stops chat drafts from being restored multiple times, preventing accidental overwrites. It improves the reliability of the chat input editor and enhances user experience.

Health Assessment

• • Rapid resolution with minimal changes; no blockers or extensive rework required.

Summary

Add Gemini 3.5 Flash support to LobeHub provider

Adds Gemini 3.5 Flash model card and pricing corrections, enabling accurate multimodal billing and structured output for LobeHub users.

Health Assessment

• • Fast turnaround with no review friction, indicating a straightforward model card update.

Summary

Remove plaintext credential tool to enhance security

Eliminates a tool that exposed credentials in LLM context, reducing data leakage risk for users.

Health Assessment

• • Fast 13.8‑hour cycle with minimal review comments; quick refactor to improve security with low risk.

Summary

Integrate hetero-agent dispatch with persistent sessions and notify protocol

Adds support for delegating tasks to external agent platforms (OpenClaw, Hermes) with persistent sessions and a notify protocol, improving agent runtime flexibility and user experience.

Health Assessment

• • Fast turnaround with minimal review, indicating a straightforward implementation that met business requirements quickly.

Summary

Add Gemini 3.5 Flash model support

Adds support for the Gemini 3.5 Flash model and updates default thinking level behavior, improving AI model selection and user experience.

Health Assessment

• • Fast cycle time and immediate AI review comments led to minimal rework; the PR was merged within 5.5 hours with no human review rounds.

Summary

Persist unsent chat drafts across tab switches

Ensures user input is retained when switching tabs or reloading, improving user experience and reducing frustration.

Health Assessment

• • Fast cycle time and minimal review comments indicate smooth integration and low risk of regressions.

Summary

Fix truncated IM responses in heterogeneous agent

Ensures IM users receive full assistant replies by using the correct message ID, improving user experience.

Health Assessment

• • Quick, single-commit fix with no tests and minimal code changes; low risk and fast turnaround.

Summary

Add prompt=consent to OIDC authorization URL

Fixes missing refresh token by forcing consent screen on each login, ensuring users can stay authenticated without re‑authenticating.

Health Assessment

• • Quick fix with minimal code changes, no rework, fast review and merge.

Summary

docs: add ph #1 badge

Adds a badge to the documentation README files.

Health Assessment

• • PR merged immediately with no review, minimal changes.

Summary

Fix App Nap dropping gateway WebSocket

Prevents gateway WebSocket disconnects during macOS display sleep by using Electron powerSaveBlocker, ensuring devices stay online without manual wake.

Health Assessment

• • Fast turnaround with minimal changes; AI review provided concise guidance and helped finalize the implementation quickly.

Summary

Refactor agent settings: consolidate tabs, drop auto-topic

Streamlines agent configuration by merging tabs into a single popover and removing unused auto-topic feature, improving UI clarity and reducing maintenance. This reduces code complexity and enhances user experience.

Health Assessment

• • The PR involved extensive refactoring across many files with rapid iteration and AI-assisted review, indicating a complex change that required careful coordination.

Summary

Fix desktop visual media URL handling

Corrects how desktop local media URLs are processed, preventing SSRF issues and ensuring proper fallback behavior for visual analysis.

Health Assessment

• • Rapid resolution with minimal review; AI review comments helped identify SSRF safety and MIME type validation.

Summary

Fix onboarding IM integration

Corrects messaging integration for onboarding, improving user experience and reducing onboarding friction.

Health Assessment

• • Rapid iteration with minimal review, large code changes, but quick merge indicates high confidence in the fix.

Summary

Refactor onboarding to single profession question

Simplifies onboarding by focusing on profession, speeding up agent selection and reducing user friction.

Health Assessment

• • Fast cycle time and minimal rework indicate a smooth, low-risk refactor with automated review support.

Summary

Fix agent signal receipt anchoring

Ensures agent signal receipts are anchored to stable message IDs, preventing drift and improving reliability of message tracking.

Health Assessment

• • Rapid review and merge within 5.5 hours indicates efficient process; large code changes but minimal review rounds suggest confidence in quality.

Summary

Optimistic interests update and clickable auth logo

Adds instant UI feedback for profile interests edits and makes auth page logo navigable to home, improving user experience.

Health Assessment

• • Rapid review and merge with minimal iterations indicates a smooth process.

Summary

Fix inline type filter promotion in command menu

Ensures type-based search filters are correctly parsed, improving search accuracy and user experience in the command menu.

Health Assessment

• • Quick fix with minimal changes, immediate merge, low review friction.

Summary

Fix home sidebar layout alignment

Aligns the home sidebar layout for consistent UI and improves navigation grouping, enhancing user experience.

Health Assessment

• • The PR was reviewed and merged within 4 hours with minimal back‑and‑forth, indicating a straightforward, low‑risk change.

Summary

Add May 19 weekly changelog

Adds a new weekly changelog entry for May 19, documenting recent product updates and improvements for both English and Chinese users.

Health Assessment

• • Fast cycle time and minimal review friction indicate a straightforward documentation update.

Summary

Add chat cost estimate support

Introduces cost estimation utilities and user settings for chat usage, enabling users to see and control chat cost predictions.

Health Assessment

• • Rapid AI‑assisted review led to a large, complex change with minimal human review time.

Summary

Add Market Agent Picker to onboarding flow

Introduces a new step in the classic onboarding process that lets users select agents from the marketplace, enhancing personalization and engagement.

Health Assessment

• • Fast cycle time (2.2h) and single review round indicate high confidence; AI review tools were used to validate the changes.

Summary

Fix agent builder panel auto‑open after prompt

Ensures users see the Agent Builder panel immediately after creating an agent via prompt, improving workflow efficiency.

Health Assessment

• • Fast cycle time (1.8h) and immediate review indicate a smooth process; minimal rework and automated review bots contributed to quick resolution.

Summary

Fix 4K resolution drop when aspect ratio auto

Ensures image generation respects user‑selected resolution even when aspect ratio is left at auto, improving image quality for nano‑banana users.

Health Assessment

• • Quick fix with minimal rework; AI‑assisted review helped catch edge cases and validate test updates.

Summary

Skip branded provider LLM retries

Avoids unnecessary retry loop for branded provider, improving performance.

Health Assessment

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

Summary

Fix IM bot-callback webhook for heterogeneous agents

Ensures IM bots receive final responses when using heterogeneous agents by storing and delivering completion webhooks, improving reliability.

Health Assessment

• • Quick resolution with minimal iterations, AI-assisted review expedited the process.

Summary

Fix stderr handling in error output

Ensures shell command errors are surfaced correctly, improving debugging and user feedback.

Health Assessment

• • Quick fix with minimal changes, reviewed immediately, no back‑and‑forth.

Summary

Fix hidden matching for dot-prefixed glob patterns

Ensures file search functions correctly handle hidden directories and files, improving reliability for users.

Health Assessment

• • Quick AI-assisted fix with minimal rework, fast review cycle, and no major risk to production.

Summary

Fix unsupported file parser rejection

Prevents unsupported file types from being processed, improving reliability and user experience by rejecting invalid uploads early and providing clear error feedback.

Health Assessment

• • The PR was merged within 0.9 hours with only one review comment, indicating a quick, low-friction fix. The large code change was well-contained and did not trigger extensive discussion.

Summary

Fix Anthropic-compatible base URL normalization

Ensures Anthropic-compatible providers correctly construct SDK URLs, preventing 404 errors and improving reliability for gateway routes.

Health Assessment

• • Fast cycle time of 1.9 hours, single review with minimal comments, and only one rework commit indicate a smooth, low-risk PR process.

Summary

Streamline issue triage to core labels

Simplifies issue labeling to 1-3 core business labels, reducing clutter and improving triage efficiency.

Health Assessment

• • PR merged immediately after review with minimal changes, indicating low friction and quick approval.

Summary

Optimize chat bootstrap persistence for performance

Improves server‑side chat initialization speed by reducing database queries and consolidating message creation, leading to faster user experience. Adds timing instrumentation for better monitoring.

Health Assessment

• • Fast review and merge cycle with minimal back‑and‑forth, indicating a well‑tested and low‑risk performance tweak.

Summary

Standardize messenger platform preposition copy

Ensures consistent wording across UI, bot messages, and docs, improving user clarity.

Health Assessment

• • Quick turnaround with minimal rework; single review comment; fast merge.

Summary

Update subscription budget recovery copy

Adds localized copy for budget recovery upgrade action, improving user experience across languages.

Health Assessment

• • The PR was reviewed and merged quickly, but the large number of lines and files indicates a substantial change to the i18n assets, which could affect many locales.

Summary

Remove compromised actions-cool/issues-helper action

Eliminates a supply chain attack vector by removing the compromised action and pinning to a safe commit, ensuring CI/CD security.

Health Assessment

• • Rapid response to supply chain attack with minimal changes and quick review, demonstrating strong security posture.

Summary

Translate non-English comments to English

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

Health Assessment

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

Summary

Fix markdown preview scrollability in LocalFile portal

Ensures long markdown files in the portal can be scrolled, improving user experience when viewing skill documentation.

Health Assessment

• • Quick, single‑commit fix with no rework, reviewed and merged within 21.5 hours.

Summary

Restore home nav for task workspace

Fixes sidebar navigation so the home sidebar appears on the /tasks route, ensuring the correct active state and updated translations.

Health Assessment

• • Fast turnaround with a single review and minimal changes, indicating low complexity and risk.

Summary

Refactor onboarding chat input feature switches

Improves onboarding chat input by grouping feature switches for completion, mentions, and slash commands, ensuring consistent behavior and easier maintenance.

Health Assessment

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

Summary

Fix agent context preservation in task routes

Ensures task navigation retains agent context, preventing context loss and improving user experience when accessing tasks.

Health Assessment

• • Rapid review and merge with minimal back‑and‑forth indicates high confidence; the change is large but well‑tested and focused on a critical navigation bug.