Pull Request Explorer

Summary

Fix element refs for browser secret capture

This fix resolves a critical bug that prevented automated browser workflows from capturing secrets when using element references, ensuring reliable credential collection for integrations like Slack. By implementing proper element value retrieval, the platform now reliably stores secrets, improving developer experience and security compliance.

Health Assessment

• • Fast cycle time of 3.5 hours with minimal review indicates a straightforward bug fix with low complexity.

Summary

Improve external secrets provider failure logs

Enhances error handling for external secrets providers, improving observability and reliability.

Health Assessment

• • The PR had a long cycle time but a rapid initial review, yet many commits after the first review indicate iterative refinement and potential friction. The large number of lines added and removed increases risk for regressions.

Summary

Polish instance AI MCP connections UI

Improves the user interface for instance AI MCP connections, adding better styling, credential handling, and navigation to enhance user experience.

Health Assessment

• • Long cycle time with multiple review rounds and significant rework indicates high friction and risk.

Summary

Remove unused references in Tournament

Clean up package configuration to eliminate dead dependencies, improving dev container startup reliability.

Health Assessment

• • PR merged within 20 hours with a single review comment, indicating a straightforward cleanup with minimal risk.

Summary

Resilient ChatHub E2E assertions against LLM drift

Improves reliability of ChatHub end‑to‑end tests by replacing brittle text assertions with structure‑based checks, ensuring consistent test results across model updates.

Health Assessment

• • Fast review and merge with minimal changes; low risk and high confidence in test stability.

Summary

Enable restoreMocks by default in shared vitest configs

This change sets restoreMocks to true globally, ensuring test spies are reset before each test. It removes redundant per-package overrides, simplifying maintenance and preventing hidden test failures.

Health Assessment

• • Quick review and merge with minimal changes; the PR was straightforward and had no blockers.

Summary

Add async and partial publication UX

Improves workflow publication experience by handling asynchronous and partial states, reducing false failures and improving user confidence.

Health Assessment

• • PR involved extensive changes across frontend and backend, with many tests added; review took almost 48 hours and required multiple commits, indicating moderate complexity and potential risk.

Summary

Release 2.29.5

This PR publishes version 2.29.5, bumping package versions and updating the changelog.

Health Assessment

• • Automated release PR merged instantly with no review comments, indicating a routine maintenance update.

Summary

Fix saving tools added via agent tools modal

Ensures agent tools added through the modal are correctly saved, preventing autosave errors and improving reliability of agent configuration.

Health Assessment

• • Rapid backport with minimal changes and immediate approval indicates low complexity and risk.

Summary

Add script to seed example instance

Provides a script to populate an n8n instance with example workflows, enabling quick setup for testing and demos.

Health Assessment

• • Fast cycle time (4.9h) and minimal review rounds indicate a smooth process; AI review helped surface issues early.

Summary

Add new Tree component to design system

Adds a modern tree component using Reka UI primitives to improve maintainability, consistency, performance, and accessibility in the design system.

Health Assessment

• • The PR introduced a substantial new component with many changes, but the review process was quick and the AI review flagged a single issue. The team made several iterative commits to address feedback, resulting in moderate review friction and a large scope.

Summary

Enable filtering crashed executions via public API

Allows public API consumers to filter executions by actual crashed status, improving visibility and debugging.

Health Assessment

• • Fast review with a single comment, minimal changes, low risk.

Summary

Track Instance AI prompt suggestion exposure

Adds an exposure event for Instance AI personalized prompt suggestions, enabling PostHog analysis and ensuring accurate tracking in the editor UI.

Health Assessment

• • Rapid backport with minimal changes and quick review, indicating low complexity and high confidence.

Summary

Minor updates to custom instance roles

Fixes critical bugs in role management, preventing errors for users with limited permissions and ensuring UI stability for long role names, improving reliability and user experience.

Health Assessment

• • The PR was reviewed and merged within 1.2 hours, indicating a smooth, low‑friction process with minimal scope and no significant blockers.

Summary

Fix Instance AI prompt caching across resumes

Improves cache hit rate for resumed runs, reducing reprocessing overhead and saving compute costs.

Health Assessment

• • Rapid backport with minimal changes, quick review, no blockers.

Summary

Parallelize MCP workflow evals across lanes

Enables parallel evaluation of MCP workflows, cutting CI time and improving resource utilization for faster feature delivery.

Health Assessment

• • Fast cycle and review with minimal rework, but large code changes suggest moderate risk of unforeseen issues.

Summary

Track Instance AI Prompt Suggestion Exposure

Adds a dedicated exposure event for Instance AI personalized prompt suggestions, enabling PostHog analysis and ensuring accurate tracking across UI states.

Health Assessment

• • Quick review and merge, minimal rework, straightforward change.

Summary

Fix Instance AI prompt caching across resumes

Improves cache hit rate for multi‑turn AI workflows, reducing token usage and cost while maintaining consistent behavior.

Health Assessment

• • Fast turnaround with a single review and minimal rework; the change is small to medium in scope and does not introduce new functionality or risk.

Summary

Hide "Needs first setup" badge for private credentials

Ensures that private credentials do not display the "Needs first setup" badge, improving UI clarity and preventing confusion for users.

Health Assessment

• • Quick review with a single comment, minimal code changes, and AI assistance led to a fast, low‑risk resolution.

Summary

Surface underlying cause on OAuth callback errors

Improves diagnostics for OAuth credential failures by exposing root cause and logging, enabling faster resolution of private credential issues.

Health Assessment

• • Quick review and single commit indicate a straightforward fix with minimal friction.

Summary

Block manual project member changes when roles managed

Ensures SSO‑managed project roles are authoritative, preventing manual overrides that could cause permission inconsistencies for users.

Health Assessment

• • Fast initial review (0.1h) but required three subsequent commits to address type‑check and test failures, indicating moderate rework on a large codebase.

Summary

Scope Instance AI evals dispatcher to PR gating

Clarifies CI workflow roles, excludes fork PRs, and improves maintainability and security of AI evaluation gating.

Health Assessment

• • Fast review and merge with minimal changes; clear documentation and role separation reduce future friction.

Summary

Preserve Instance AI output for interrupted runs

Ensures users retain AI-generated work when stopping or timing out a run, preventing data loss and improving continuity of subsequent interactions.

Health Assessment

• • The PR required several commits to address persistence and UI updates, but the review was rapid and comments were minimal, indicating an efficient resolution process.

Summary

Release 2.28.6

Patch release addressing core and editor bugs, enhancing stability and AI gateway functionality.

Health Assessment

• • Fast release with minimal changes and no review delays.

Summary

Release 2.29.4

Bug fixes and new AI assistant features for workflow automation, improving reliability and expanding node capabilities.

Health Assessment

• • Merged within 18 minutes with no review friction, indicating a streamlined release process.

Summary

Add n8n-package-exported logstreaming event

Enables audit logging of package exports for SIEM integration, improving traceability of who exported what and when.

Health Assessment

• • Fast review and minimal rework led to a quick 18‑hour cycle time, indicating a smooth integration of the new audit event.

Summary

Stabilize Instance AI workflow execution tests

Improve reliability of Instance AI workflow execution tests by re-recording expectations and addressing replay issues.

Health Assessment

• • The PR has a large scope with many changed files, but the review process was relatively quick and smooth.

Summary

Fix saving tools added via agent tools modal

Fixes autosave error caused by mismatched modal payload, ensuring added tools are correctly saved and displayed.

Health Assessment

• • Quick fix with minimal changes, approved immediately, no rework needed.

Summary

Enhance Agent Knowledge Base Retrieval

Improves the agent’s knowledge base retrieval, boosting accuracy and response speed for end users.

Health Assessment

• • Merged within minutes with a single commit and minimal changes, indicating a straightforward, low‑risk improvement.

Summary

Fix credit banner spacing in Instance AI

Corrects visual layout of the credit warning banner in Instance AI views, ensuring consistent appearance across UI components.

Health Assessment

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

Summary

Reconcile backport labels on minor releases

Automates label adjustments for backported PRs during minor releases, ensuring correct labeling and reducing manual effort.

Health Assessment

• • Fast review and merge with minimal iterations; AI assistance streamlined changes.

Summary

Prevent duplicate zod instances that break npm installs

Fixes npm install crashes by ensuring a single shared zod instance across packages, restoring peer dependency and adding guard scripts to prevent future regressions.

Health Assessment

• • Quick backport with minimal code changes, low review friction, and no significant risk to production.

Summary

Fix eval backend tracing and model input in CI

Adds LangSmith tracing to eval CI lanes, enabling thread visibility. Also introduces a model input for builder experiments, allowing flexible model selection.

Health Assessment

• • Fast review and merge with minimal changes, indicating a straightforward bugfix with clear test coverage.

Summary

Fix AI Gateway hints for tool nodes

Ensures AI Gateway notices appear for nodes used as tools, improving user experience and consistency.

Health Assessment

• • Quick backport with minimal changes, no blockers, fast review.

Summary

Prevent duplicate zod instances that break npm installs

Fixes npm install crashes by ensuring a single shared zod instance across packages, preventing future dependency conflicts.

Health Assessment

• • Backport resolved a critical npm install crash with minimal changes and AI-assisted code generation, leading to a fast merge.

Summary

Fix AI Gateway hints for tool nodes

Ensures AI Gateway notices appear for nodes used as tools, improving user experience and consistency across workflows.

Health Assessment

• • Quick backport with minimal changes and no review delays, indicating a straightforward bug fix.

Summary

Fix Instance AI checkpoint growth and expiry

Ensures database cleanup and secure confirmation handling, preventing stale data and potential misuse of Instance AI features.

Health Assessment

• • Rapid 1.6‑hour cycle with minimal review comments indicates a straightforward, low‑risk bug fix. AI‑generated code was used, but the changes were small and well‑tested.

Summary

Add model input to Instance AI eval workflow

Enables experimentation with different AI models by adding a free‑form model input to the Instance AI evaluation workflow, allowing comparisons without code changes.

Health Assessment

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

Summary

Hide custom scopes for managed OAuth credentials

Improves UI by hiding unnecessary fields for managed OAuth credentials, reducing confusion and potential errors for users.

Health Assessment

• • Fast turnaround with AI-assisted review, minimal rework, and clear business impact.

Summary

Fix AI Gateway hints for tool nodes

Ensures AI Gateway hints display correctly for nodes used as tools, improving user experience and preventing missed warnings.

Health Assessment

• • Quick review and merge with minimal changes; no blockers.

Summary

Prevent duplicate zod instances causing npm install crashes

Fixes npm install crashes by ensuring a single shared zod instance across packages, restoring stability for users.

Health Assessment

• • Quick resolution with minimal rework; AI-generated summary helped streamline review.

Summary

Add durable scheduler storage adapter

Adds a persistent storage layer for the scheduler, enabling reliable job execution and future scalability of the scheduling engine.

Health Assessment

• • The PR was reviewed and approved within 0.2 hours, indicating a smooth review process, but the large code change (over 500 lines) introduces higher risk for integration and testing complexity.

Summary

Add folder export to n8n packages

Enables users to export folder structures as .n8np packages via CLI and API, facilitating sharing and backup of workflow configurations.

Health Assessment

• • The PR had a long cycle time with a single review comment, indicating a slow review process but ultimately merged after a single iteration.

Summary

Fix deterministic ordering of data table rows

Ensures pagination returns consistent rows by adding deterministic ORDER BY, preventing duplicates and missing rows on large tables.

Health Assessment

• • Quick review and merge, minimal changes, low risk.

Summary

Expand create-instance-ai-eval skill documentation

Enhances documentation for the create-instance-ai-eval skill, adding progressive-disclosure guidance, case archetypes, and run recipes, improving clarity for users and developers.

Health Assessment

• • Docs update completed quickly with minimal review, indicating efficient documentation process.

Summary

Add prompt caching for Anthropic and OpenAI agents

Enables efficient reuse of prompts across agent calls, reducing API usage and cost for Anthropic and OpenAI models.

Health Assessment

• • The PR had a moderate cycle time but involved many commits and a large code change, indicating significant effort. The single review round and low comment count suggest a smooth review process, but the large scope raises potential risk.

Summary

Set default pnpm build in CI action

Ensures CI workflow uses the correct pnpm build command, improving cache efficiency and preventing build mismatches.

Health Assessment

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

Summary

Support community node type definitions in AI Assistant

Instance AI can now resolve type definitions for community and custom nodes, expanding AI assistance to a broader set of workflows. This reduces manual configuration and improves user experience.

Health Assessment

• • Fast cycle time and immediate approval indicate high confidence in the change; minimal review friction and a single commit suggest a straightforward implementation.

Summary

Refactor Prometheus metric query extraction

Streamlines metric collection by centralizing caching logic, reducing redundant queries and improving performance for Prometheus metrics.

Health Assessment

• • Fast cycle time and minimal review rounds indicate a smooth, low‑risk refactor with a single commit and clear AI‑generated documentation.

Summary

Fix manual execution offloading in queue mode

Ensures manual executions are correctly persisted and processed when offloaded to workers, preventing errors and enabling manual runs without triggers.

Health Assessment

• • Quick fix with minimal changes, single review, no blockers.