Pull Request Explorer

Summary

Snapshot database tool input schemas for MCP

Adds comprehensive snapshot tests for database CRUD tool schemas, ensuring future refactor to JSON Schema passes without breaking existing behavior.

Health Assessment

• • Fast review and merge within 1.2 hours, minimal rework, large snapshot test ensures future refactor stability.

Summary

Improve API performance by excluding timeline activities

Reduces API payload size and GC pauses, improving response times and user experience.

Health Assessment

• • Fast review and merge with minimal rework; AI review bots provided quick feedback; performance improvement with low risk to existing functionality.

Summary

Add observability for MCP method and tool names

Enhances logging and tracing of JSON‑RPC method and tool names to enable faster diagnosis of slow API calls.

Health Assessment

• • Fast review cycle (0.1h) and minimal changes (174 lines) indicate low complexity and high confidence in the implementation.

Summary

Improve Markdown Parsing Performance

Replaces synchronous DOM parsing with a fast lexer, cutting conversion time from seconds to milliseconds for large markdown notes and eliminating API pod stalls.

Health Assessment

• • Rapid AI-assisted review and multiple small commits enabled a quick merge, delivering significant performance gains for API pods.

Summary

Fix calendar sync to skip unchanged event updates

Reduces unnecessary database writes during calendar sync by skipping updates for unchanged events and participants, improving performance and reducing load.

Health Assessment

• • Rapid iteration with AI-assisted review enabled quick resolution; large code changes but minimal review friction.

Summary

Migrate phone country picker to Twenty UI

Adds a modern phone country picker component to the Twenty UI library, enabling users to search countries by name or calling code with flag icons, improving data entry accuracy and user experience.

Health Assessment

• • Rapid iteration with AI-assisted review led to quick resolution and a comprehensive feature addition.

Summary

Memoize credentialed CORS origins for performance

Caches parsed CORS origins to cut URL parsing overhead per request, improving server response time and memory usage.

Health Assessment

• • Merged within an hour with a single review, indicating a straightforward performance tweak with minimal risk.

Summary

Let buttons inside field inputs handle Enter

Field inputs now correctly handle the Enter key when interacting with embedded buttons or links, preventing unintended form submissions and improving user experience.

Health Assessment

• • Rapid review and merge with minimal iterations indicates low complexity and high confidence in the change.

Summary

Preserve first keystroke in dropdown search

Ensures user input is not lost when opening a dropdown, improving UX and preventing frustration from lost search queries.

Health Assessment

• • Fast turnaround with minimal rework; auto-approval and AI-assisted review streamlined the process.

Summary

Migrate currency picker to Twenty UI

Integrates the currency picker into the Twenty UI component library, enhancing consistency and usability across the platform.

Health Assessment

• • Rapid AI-assisted review enabled approval within 5.5 hours, demonstrating efficient collaboration and minimal friction.

Summary

Improve workspace cache packing latency

Extends the idle threshold for cache packing to 10 minutes, reducing unnecessary packing and event‑loop delays while slightly increasing memory usage per pod.

Health Assessment

• • Fast, single‑commit PR with auto‑approval and no significant rework; minimal risk and quick turnaround.

Summary

Fix GraphQLPlayground story worker load errors

Resolve intermittent test failures caused by Monaco worker load errors in Storybook, improving test reliability and developer experience.

Health Assessment

• • Quick fix with minimal changes, auto‑approved, no significant rework.

Summary

Align visibility cell and add back arrow

Fixes UI alignment for core workflow visibility and restores navigation in the versions panel, improving user experience and reducing confusion.

Health Assessment

• • Rapid review and minimal changes indicate a low‑risk, high‑confidence fix. AI review helped catch UI inconsistencies early.

Summary

Expose universal identifier validation across services

Centralizes identifier validation logic, eliminating duplication and ensuring consistent behavior across the server, SDK, and shared utilities, improving maintainability and reducing future bugs.

Health Assessment

• • Fast cycle time (1.4h) and immediate review approval indicate low friction; single review round and minimal comments suggest a straightforward refactor with minimal risk.

Summary

Stop reporting inactive workflow triggers to Sentry

This change stops Sentry from logging dropped workflow triggers that are expected when a workflow version is archived or deactivated, reducing noise and improving alert relevance.

Health Assessment

• • Quick fix with minimal code changes, no impact on functionality, and no risk to production.

Summary

Improve cache rebuild performance by reducing lookups

This change eliminates quadratic lookup construction during cache rebuilds, cutting CPU usage and reducing event-loop stalls. It improves API responsiveness and prevents pod restarts caused by liveness probe failures.

Health Assessment

• • Fast review and merge with minimal changes, indicating high confidence in performance gains and low risk to production stability.

Summary

Update snapshot for agentThread auth coverage

Ensures agentThread entity snapshots reflect new authentication coverage, improving test reliability.

Health Assessment

• • Quick snapshot update with no review needed, minimal risk.

Summary

Keep agent steps in one group and surface reasoning summaries

Improves chat UI by grouping record widgets, hiding empty reasoning, and enabling optional OpenAI reasoning summaries, reducing clutter and improving user experience.

Health Assessment

• • Fast cycle time and minimal review rounds indicate a smooth process; AI-assisted review helped surface issues quickly.

Summary

Snapshot AuthPrincipalGuard and application target per endpoint

Captures guard and target configurations for all endpoints, enabling quick detection of permission changes and reducing regression risk.

Health Assessment

• • Fast review cycle with AI-assisted summary; large snapshot test added with minimal code changes; quick merge within 22 hours.

Summary

Improve workspace member cache mapping

Optimizes hot‑path queries by caching workspace member lookups, reducing database load for AI tool calls, role resolution, and record sharing, leading to faster response times.

Health Assessment

• • Fast cycle time and minimal review comments indicate smooth integration; performance gains achieved with minimal code churn.

Summary

Refactor ORM lean write path into performMutation

Streamlines mutation logic, reducing code duplication and potential bugs without affecting functionality.

Health Assessment

• • Quick turnaround with minimal review, indicating a straightforward refactor.

Summary

Improve permission flag resolution via workspace cache

Reduces database queries for permission checks, improving performance and reliability for users and API keys.

Health Assessment

• • Rapid 4‑hour cycle with a single review round and minimal comments, indicating a well‑structured change; the large scope and many line changes suggest moderate risk but the quick review suggests confidence in the implementation.

Summary

Improve role resolver performance and add getRole query

This PR optimizes role resolution by batching database calls and introduces a dedicated getRole query, reducing latency and resource usage for role-related operations, benefiting admins and users with faster UI responses.

Health Assessment

• • The PR was reviewed and merged within 3 hours, indicating efficient collaboration, but the large code changes and performance focus suggest a high-impact update that required careful testing.

Summary

Reshape record sharing API and share panel

Enhances record sharing logic, introduces a new API, eliminates polling, and improves permission handling for a smoother user experience.

Health Assessment

• • Rapid iteration with many commits and a quick review, but the large scope and breaking API changes elevate risk.

Summary

Preserve OAuth-only caller reach in application targeting

Ensures OAuth-only clients can correctly access applications they are authorized to reach, improving security and functionality.

Health Assessment

• • Quick review and merge with minimal rework, indicating a straightforward bug fix with low risk.

Summary

Bump version to 2.46.0

Updates package versions and internal constants to new release, ensuring consistency across client, UI, and server components.

Health Assessment

• • Fast cycle time and minimal changes indicate low risk and efficient process.

Summary

Refactor ORM: Remove Query Builder Mutations

Eliminates unchecked mutation paths in the ORM, tightening permission checks and reducing the risk of unauthorized data changes while maintaining performance for system‑level operations.

Health Assessment

• • The PR was merged in 3.5 hours with minimal review friction, largely due to AI‑assisted review comments and a focused refactor that avoided new feature risk.

Summary

Add test to wait for recomputed statuses

Ensures workflow status polling waits for recomputed values, preventing flaky CI failures.

Health Assessment

• • Fast review and merge with minimal rework; test stabilizes workflow status polling.

Summary

Log structured line per failed AI chat turn

Adds detailed failure logging for AI chat turns, improving debugging and monitoring.

Health Assessment

• • Rapid review and merge within 3 hours, minimal rework, AI review tool assisted.

Summary

Refactor permissions to single source of truth

Simplifies permission handling across the app, improving consistency and performance, reducing bugs and rework.

Health Assessment

• • Rapid iteration with multiple small commits indicates active refactoring; high scope and many changes increase risk but quick review cycle mitigates delay.

Summary

Fix navigate_app to use explicit IDs

Ensures AI navigation respects permissions, preventing data leaks and improving security for users.

Health Assessment

• • Rapid 2.3‑hour cycle with multiple AI‑reviewed iterations indicates a focused security fix, but the large code churn and many review rounds suggest elevated risk.

Summary

Fix frontend wildcard Accept header handling

Ensures frontend requests without Accept headers or with wildcard are served correctly, preventing 404s for health checks and CLI clients.

Health Assessment

• • Fast cycle time of 0.8h, single auto‑approved review, minimal code changes, low risk of regressions.

Summary

Pin adm-zip to 0.6.1 for all consumers

This update addresses five critical security advisories by upgrading adm-zip to 0.6.1 across all consumers, ensuring protection against decompression attacks and privilege escalation.

Health Assessment

• • Rapid review and merge with minimal changes; AI review confirms no issues.

Summary

Consolidate validation rule preview into single card

Improves UI consistency by merging preview, editor, and suggestion panels into one card, enhancing user experience for validation rule configuration.

Health Assessment

• • Fast cycle time and auto‑approved reviews indicate low complexity and minimal friction; the change is small to medium in scope and poses no significant risk.

Summary

Add validation rule suggestion UI details

Enhances the validation rule form UI to match the design, improving user clarity and consistency.

Health Assessment

• • Fast 11.7‑hour cycle with minimal review friction; small to medium scope and auto‑approvals indicate low complexity and risk.

Summary

Strip NUL bytes from AI chat message parts

This change removes problematic NUL bytes from AI chat outputs, preventing Postgres errors. It ensures users receive complete answers without interruption.

Health Assessment

• • Fast cycle time and minimal review friction indicate efficient handling of a small bugfix.

Summary

Add MenuItem suggestion rows to validation rule editor

Enhances the validation rule editor by replacing custom divs with native MenuItem components, improving usability and consistency with the design, and reducing user confusion.

Health Assessment

• • The PR was reviewed and merged within 11 hours, with minimal changes and no rework, indicating a smooth, low‑risk process.

Summary

Fix row-level permission handling in query builder

Corrects query builder logic so row-level permissions are always enforced, eliminating potential data exposure when OR filters are applied.

Health Assessment

• • Fast review and merge with minimal rework, indicating a straightforward bug fix.

Summary

Security bump of decompress to 11.1.4

Updates the decompress package to a patched version, eliminating a critical path traversal vulnerability that could affect builds.

Health Assessment

• • Fast turnaround with minimal changes; AI review flagged no issues; security fix applied promptly.

Summary

Remove dead and duplicated permission code

Clean up unused permission logic to reduce technical debt and improve maintainability without affecting user-facing behavior.

Health Assessment

• • Rapid review and merge with minimal changes, indicating straightforward cleanup.

Summary

Render field chips with shared Chip component

Aligns validation rule form with the Configure field design, improving UI consistency and reusability for users.

Health Assessment

• • The PR was reviewed and auto-approved within 10 hours, indicating low complexity and quick turnaround. The small number of files and moderate line changes suggest a low-risk, high-value UI improvement.

Summary

Fix role save logic and permissions fallback

Corrects double navigation on role creation, ensures proper saving state handling, and fixes permissions fallback to use correct shape, improving user experience and data integrity.

Health Assessment

• • Fast cycle with minimal review, but large code changes across many files, indicating significant refactor and bug fixes.

Summary

Bump piscina dependency to address critical vulnerability

Fixes a critical prototype‑pollution vulnerability by updating piscina to 4.9.4, ensuring a secure build environment. This update prevents potential code execution via unsafe ThreadPool options.

Health Assessment

• • Rapid review and auto‑approval indicate low complexity and minimal risk.