Pull Request Explorer

Summary

Stabilise Codecov uploads and tighten coverage scope

Improves reliability of CI coverage metrics, reduces false positives, and provides clearer reporting for AI and backend components.

Health Assessment

• • Quick review and merge, minimal changes, stable CI improvements.

Summary

Release 1.123.46

Patch release with bug fixes, including requiring admin push for legacy environment migration.

Health Assessment

• • Quick release with minimal changes and no issues found by the reviewer.

Summary

Generate release SBOM from shipped artefact

Creates a more accurate SBOM by scanning the shipped artifact instead of the source tree, reducing false positives in license audits and cutting component count from 6,065 to 1,399.

Health Assessment

• • Fast review and merge with minimal rework, but large lockfile changes added many lines of configuration.

Summary

Enable include-hidden-files on shard artifact upload

Fixes E2E coverage uploads by ensuring dotfiles are included, restoring coverage reports and Codecov flag updates.

Health Assessment

• • The PR required a single review after a 42‑hour delay, with minimal rework and a small scope, indicating a straightforward CI fix.

Summary

Fix scheduled poll expression isolate acquisition

Ensures scheduled polling triggers acquire an expression isolate, preventing runtime errors under the vm expression engine and restoring reliable trigger execution.

Health Assessment

• • The PR required a backport and involved a single commit after review, indicating moderate iteration but a straightforward bug fix.

Summary

Route private OAuth2 credentials through dynamic-credential origin

Enhances OAuth2 authentication by routing private credentials through dynamic-credential origin, improving security and user experience.

Health Assessment

• • The PR had multiple review rounds and required several commits to address concerns, indicating some complexity and potential risk.

Summary

Remove 'Continue Chat' for close button in Agents

Updates the Agent session timeline header so users can close the session detail view instead of continuing the chat from that page, improving navigation clarity.

Health Assessment

• • The PR had a long cycle time due to a delayed review and a single rework commit, but the change was small and straightforward, resulting in low risk.

Summary

Remove 'no more sessions' message from Agent sessions list

This change removes the redundant 'no more sessions' message from the Agent sessions list, streamlining the UI and reducing confusion for users.

Health Assessment

• • The PR was reviewed by an AI bot with no issues, and merged after a single review, indicating a straightforward change with minimal friction.

Summary

Add dropEnumCheck helper to migration DSL

Provides a DSL method to drop enum CHECK constraints, simplifying migration scripts and reducing manual query construction.

Health Assessment

• • Fast 1.2‑hour cycle with AI review enabled quick approval and minimal rework.

Summary

Streamline fix-with-AI offer state and reporting

Enhances the AI fix feature by streamlining state handling and reporting, cutting runtime overhead. This leads to faster user interactions and easier maintenance.

Health Assessment

• • Quick review and merge with minimal changes indicates a smooth refactor with low risk.

Summary

Add episodic memory knowledge tab

Introduces an episodic memory knowledge tab in the agent builder, enabling users to view, search, and manage memory entries, improving transparency and control over AI behavior.

Health Assessment

• • The PR received a quick initial review but required several subsequent commits to address issues, indicating moderate iteration but overall efficient resolution.

Summary

Fix Git node clone operation to use target path

Ensures Git repositories are cloned into the specified target path, preventing incorrect cloning into the current directory and improving reliability for users.

Health Assessment

• • Quick AI‑assisted fix with minimal changes, resolved in under 18 hours, demonstrating efficient issue resolution.

Summary

ci: Relax Instance AI discovery evals

Allow partial threshold misses in PR CI for Instance AI discovery evals, while maintaining strict local behavior.

Health Assessment

• • The PR was reviewed and approved quickly, with no significant issues found.

Summary

Harden setup-nodejs to prevent silent Node-version fall-through

Improves CI reliability by ensuring Node version matches expectations, reducing flaky builds and saving runner minutes. This reduces debugging time and costs for developers.

Health Assessment

• • Fast review and merge with minimal iterations; CI improvement with low risk.

Summary

Skip coverage collection on merge_group runs

Reduces CI queue time by skipping unnecessary coverage uploads during merge queue runs, improving build throughput without affecting PR review experience.

Health Assessment

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

Summary

Fix migration timestamp override handling

Corrects static analysis rule to allow dialect-specific migration overrides with identical timestamps, preventing false positives that could block CI pipelines and deployment. This ensures smoother CI runs and accurate migration ordering for multi-dialect deployments.

Health Assessment

• • Quick fix with minimal changes, reviewed promptly, no rework needed.

Summary

Fix manual evaluation run loop

Ensures all dataset rows are processed when a workflow is manually executed, improving reliability and preventing incomplete evaluations for users.

Health Assessment

• • Quick, single-commit fix with minimal code changes and no blockers, indicating a low-risk, high-confidence resolution.

Summary

Redesign Instance AI Approval Dialogs

Improves user experience by providing clearer, keyboard‑navigable approval options and consistent i18n titles, reducing confusion and streamlining AI interactions.

Health Assessment

• • The PR underwent multiple iterations and a significant number of commits after the initial review, indicating substantial rework. The large scope and UI changes suggest moderate to high risk, but the review process was relatively efficient.

Summary

Use i18n heading for insights chart granularity title

Fixes the insights dashboard chart title to use localized heading, improving user experience by displaying proper language‑specific labels.

Health Assessment

• • Quick AI‑assisted fix with minimal code changes, fast review, low risk to production.

Summary

Release 2.21.5-exp.0

Prepares the project for the next experimental release by updating package versions and the changelog, ensuring dependencies and documentation are current.

Health Assessment

• • Fast merge with no review friction, indicating a straightforward release update.

Summary

Surface SSO claims and role mapping diagnostics

Enhances visibility into SSO role mapping, enabling operators to diagnose and resolve mapping issues quickly, reducing support overhead.

Health Assessment

• • Quick review and approval with minimal iterations, indicating a straightforward change with clear intent.

Summary

Vendor uudecoder, remove uuencode dependency

Eliminates an undeclared‑license dependency by vendoring a lightweight uudecoder, improving compliance and reducing bundle size.

Health Assessment

• • Quick fix with minimal rework, AI review flagged no issues, fast cycle time indicates smooth process.

Summary

Add inbound secret module with runtime credential accessor

Introduces a secure mechanism to strip and store sensitive trigger payload fields, limiting credential exposure and enabling controlled access for downstream nodes.

Health Assessment

• • PR required multiple iterations after initial review, indicating complex changes; review time was moderate but overall cycle time was long due to extensive code modifications.

Summary

Add eval-data populator tool

Automates population of evaluation data tables, streamlining AI evaluation workflows and reducing manual data entry.

Health Assessment

• • Fast cycle time of 41.7h with a single review and no rework indicates a smooth, low-risk integration of the new eval-data orchestration tool.

Summary

Add eval setup sub-agent

Adds a background sub-agent that patches workflows for evaluation, enabling automated evaluation pipelines.

Health Assessment

• • The PR had a quick initial review but required multiple subsequent commits, indicating iterative refinement. The large code change and many files suggest a high-impact feature with moderate risk.

Summary

Migrate @n8n/syslog-client from Jest to Vitest

Aligns the syslog client test runner with the monorepo's Vitest migration, ensuring consistent testing and type safety across packages.

Health Assessment

• • Quick review and merge with minimal changes indicates a smooth process.

Summary

Add system resolver id lookup to DynamicCredentialsProxy

Exposes getSystemResolverId to allow callers to resolve the seeded SYSTEM_RESOLVER_ID without database lookups, improving performance and decoupling credential resolution logic.

Health Assessment

• • Fast cycle time and minimal review comments indicate low friction. AI assistance helped streamline the implementation.

Summary

Refactor node-data access via typed-RPC dispatcher

Internal refactor to improve security and type safety of node data access, enabling future typed RPC operations without affecting runtime performance. No user-facing impact, but sets foundation for future features.

Health Assessment

• • Rapid review and merge with minimal iterations, indicating straightforward refactor with clear test coverage.

Summary

Place caret inside expression brackets on auto-switch

Fixes caret placement when auto‑switching to the expression editor, keeping the cursor inside the brackets so users can type without extra clicks.

Health Assessment

• • Single review, quick fix, minimal changes, low risk

Summary

Split Instance AI evals into separate workflow

Cuts CI eval cost by ~90% by restricting the Instance AI workflow to run only on review approval, reducing cost without affecting build/tests.

Health Assessment

• • Rapid review and minimal rework; cost savings achieved quickly.

Summary

Lock workflow redaction settings when enforced at instance level

Adds UI lock for redaction settings when instance-level enforcement is active, preventing editors from overriding policy.

Health Assessment

• • Long review cycle with feature flag gating; moderate complexity with UI changes and tests.

Summary

Inline JWKS in OAuth2 dynamic client registration

Adds inline JWKS support for OAuth2 credentials, enabling IdPs to encrypt token responses without needing external key endpoints, improving security and flexibility for self-hosted instances.

Health Assessment

• • The PR required a long review cycle and a single comment before approval, indicating moderate complexity but no major rework. The large code additions suggest significant feature implementation.

Summary

Validate non-empty prompts in AI vendor nodes before API calls

Ensures AI vendor nodes validate prompts, preventing wasted API calls and resource usage, improving workflow reliability.

Health Assessment

• • Fast review and merge with minimal rework, clear tests confirm behavior.

Summary

Validate non-empty prompts in AI vendor nodes

Ensures workflows fail fast when prompts are empty, preventing wasted API calls and resource usage.

Health Assessment

• • Fast cycle time and minimal review iterations indicate efficient resolution of a critical validation bug across multiple AI vendor nodes.

Summary

Add SDK-aware template fetching

Ensures each instance pulls templates matching its bundled SDK version, improving compatibility and reducing errors.

Health Assessment

• • Fast merge with minimal review, indicating straightforward change.

Summary

Add Playwright LangSmith eval scaffolding

Consolidates multiple dev‑time evaluation harnesses onto Playwright with a shared LangSmith fixture and reporter, enabling unified testing and reporting for n8n's internal evals.

Health Assessment

• • Quick review and single iteration, indicating straightforward changes with minimal risk.

Summary

Populate manual user ID on webhook execution

Ensures manual webhook executions deliver real‑time data by threading the user ID through the execution path, preventing infinite spinners and data loss.

Health Assessment

• • Rapid backport with AI review, minimal code changes, fast merge, low risk

Summary

Populate manual user ID on webhook execution

Ensures manual webhook executions correctly propagate user identity, fixing data delivery issues in worker mode.

Health Assessment

• • Fast review and merge with minimal changes; issue fixed with small code addition and tests.

Summary

Allow shadowing dangerous globals as variable names in workflow SDK

Fixes validation errors when user declares local variables that shadow dangerous global names, improving developer experience and preventing false positives.

Health Assessment

• • The PR had a moderate review cycle, with a single AI‑assisted review comment, and resolved a validation bug that improved developer experience.

Summary

Scope source control data tables by project

Ensures source control data export, import, and status operations are limited to the current project context, preventing cross‑project data leakage and improving data integrity.

Health Assessment

• • Fast cycle time with minimal rework; AI‑assisted review confirms no issues.

Summary

Add AI fix action to instance AI chat on execution errors

Adds a UI card that triggers an AI agent to automatically fix execution errors in the editor, improving developer productivity by reducing manual debugging.

Health Assessment

• • Fast review and merge within 26 hours, but the change spans 730 lines across 12 files, indicating a substantial feature addition with moderate risk.

Summary

Validate non-empty prompts in AI vendor nodes before API calls

Ensures AI vendor nodes validate prompts to avoid wasted API calls and provide clear error messages, improving workflow reliability and cost efficiency.

Health Assessment

• • Fast review and merge within 3 hours indicates low friction; large code changes but minimal rework suggests stable implementation.

Summary

Fix workflow active state on external hook failure

Ensures workflows remain active when an external hook rejects publish, improving reliability for users.

Health Assessment

• • Fast review and single rework commit indicate a smooth process.

Summary

Add unpinNodes opt‑in field to eval execution

Introduces an opt‑in field for unpinning nodes during AI evals, enabling future vendor SDK interception without affecting current behavior.

Health Assessment

• • Fast cycle time with minimal review iterations; large test suite added but no runtime changes, indicating low risk to existing functionality.

Summary

Reduce agents memory usage

This PR cuts idle memory usage of the agents module by 77%, improving performance and resource efficiency for all users.

Health Assessment

• • Long review cycle and large code changes indicate significant effort; memory optimization achieved but process was slow.

Summary

Fix expression extension property lookup consistency

Ensures consistent property name handling in expression runtime, improving reliability and aligning with legacy behavior.

Health Assessment

• • Quick turnaround with minimal rework; AI review flagged no issues.

Summary

Fix dynamic property key coercion in runtime

Ensures runtime denylist checks correctly handle non-string keys, preventing potential security bypasses.

Health Assessment

• • Quick fix with minimal changes, AI-assisted review, resolved in under a day.

Summary

Refactor workflow ID usage in editor components

Updates the editor UI to use the new workflow ID injection pattern, improving consistency and reducing coupling.

Health Assessment

• • The PR required significant rework after a late review, indicating complexity and potential for bugs.
• • Long cycle time and slow review suggest bottlenecks in the review process.

Summary

Refactor VM engine jmespath resolution

Enhances security by restricting host-callable surface and preventing expression injection, improving reliability of VM engine expressions.

Health Assessment

• • Single commit with quick AI review and minimal comments indicates a smooth, low‑friction refactor.

Summary

Fix credential supportedNodes preservation in lazy loader

Ensures credential testing functions work correctly, preventing errors when retrying credentials and improving user experience in the credential modal.

Health Assessment

• • Long cycle time due to scheduling, but minimal changes and a single review indicate low complexity and low risk.