Pull Request Explorer

Summary

Add per-workspace job retention override

Superadmins can set custom job retention windows per workspace, enabling fine-grained cost control and compliance.

Health Assessment

• • PR merged within 22 hours with a single review, large code changes, AI-assisted code generation, indicating efficient collaboration.

Summary

Pin ansible interpreter to persistent path

Ensures ansible jobs run reliably by preventing intermittent interpreter‑missing errors, improving platform stability for users.

Health Assessment

• • AI‑assisted fix resolved intermittent ansible job failures with minimal changes, merging within 1.1 hours.

Summary

Fix clearer errors on auto-draft save failure

Provides detailed error messages and hover tooltips for auto-draft failures, improving user experience and reducing support overhead.

Health Assessment

• • Quick AI-assisted fixes resolved backend error handling and frontend UX in under an hour, demonstrating efficient collaboration.

Summary

Fix autosave for new SDK builder items

Ensures drafts are automatically saved when creating new scripts or flows via SDK wrappers, preventing data loss and improving user experience.

Health Assessment

• • AI-assisted code generation and comprehensive tests allowed a swift, single-commit fix with no review friction.

Summary

Fix draft autosave after AI session round-trip

Ensures user edits persist when switching between AI sessions and the workspace, preventing data loss and improving the editing workflow.

Health Assessment

• • The PR was merged within 0.7 hours with minimal changes and no review iterations, indicating a straightforward, low-risk fix.

Summary

Authorize S3 reads for logged‑in deployed app viewers

Enables logged‑in users to view and download S3 files in deployed apps using the app author’s permissions, improving data accessibility while maintaining security.

Health Assessment

• • Rapid single‑commit merge with minimal review, but large code changes across backend and frontend indicate high complexity and potential risk.

Summary

Add local-review-codex skill and CI bump

Adds a local Codex review skill that mirrors CI, enabling contributors to catch issues before pushing, and updates CI to use the latest Codex model.

Health Assessment

• • Rapid merge with minimal changes, indicating a straightforward enhancement.

Summary

Fix stale ?new_draft flag resetting editor

Prevents accidental clearing of user drafts when navigating back to an editor with a stale new_draft flag, ensuring data integrity and a smoother user experience.

Health Assessment

• • Quick resolution of a user‑facing bug with minimal code changes, demonstrating an efficient review and merge process.

Summary

Enforce read authorization for S3 signing

This fix ensures that only users with explicit read permissions can mint S3 signatures, preventing unauthorized data access and strengthening workspace security.

Health Assessment

• • Fast 5.5‑hour cycle time and minimal code changes indicate a straightforward security patch with low risk.

Summary

Fix agent-worker server processor self-healing

Ensures agent workers can complete jobs without server crashes, preventing 500 errors and job restarts, improving reliability.

Health Assessment

• • Quick AI-assisted fix resolved critical server crash issue with minimal code changes, restoring reliable job completion.

Summary

Bump Rust toolchain to 1.97.0 across CI

Keeps worker builds on the latest stable Rust, preventing lint failures and ensuring consistency.

Health Assessment

• • Rapid merge with minimal changes indicates low risk and efficient process.

Summary

Add multi-workspace MCP tokens via gateway

Enables a single MCP token to access all workspaces a user can access, reducing token proliferation and simplifying client integration. This streamlines workflow for users with many workspaces and improves security by enforcing workspace scopes.

Health Assessment

• • Rapid iteration with a single review cycle; large code changes across backend and frontend; AI-assisted description generation.

Summary

Show nested restart button for subflows in containers

Adds missing restart button for deeply nested subflows, improving user ability to restart from any node and enhancing workflow reliability.

Health Assessment

• • Fast turnaround with minimal review, indicating a straightforward bug fix.

Summary

Unblock EE CUDA build on Debian 13

Fixes build failure caused by SHA1 rejection in Debian 13 by scoping a relaxed crypto policy for NVIDIA CUDA apt repository during build, restoring CI pipeline stability.

Health Assessment

• • Quick fix with minimal changes, resolved build issue promptly, no major rework.

Summary

Show optimistic user message before fork creation

Improves user experience by displaying message and loading indicator immediately during fork creation, reducing confusion and perceived latency.

Health Assessment

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

Summary

Release version 1.754.0

This PR bumps the release to 1.754.0, updating dependencies and configuration files to prepare for the new version.

Health Assessment

• • Release was automated and merged quickly with minimal review, indicating a smooth release process.

Summary

Scope AI session flow editors to session workspace

Ensures AI session editors operate on the correct workspace, preventing data leakage and improving user experience.

Health Assessment

• • Quick resolution with a single AI-assisted commit; minimal review friction and no major regressions reported.

Summary

Skip redundant retry-chain job query

Eliminates unnecessary network requests for successful top‑level script runs, improving page load performance without affecting UI.

Health Assessment

• • Merged within 5.6 hours with no review comments, indicating a straightforward performance tweak with minimal risk.

Summary

Fix AI session open for new items

Ensures new scripts, flows, and raw apps open correctly in AI sessions, preventing not‑found errors and displaying friendly tab names, improving user experience.

Health Assessment

• • Quick fix with minimal rework; resolved within 20 hours, low complexity, no major blockers.

Summary

Propagate script timeout on perpetual restarts

Ensures perpetual scripts honor configured timeouts on restarts, preventing fallback to instance defaults.

Health Assessment

• • Quick fix with minimal changes, merged within 20 hours, indicating low complexity and high confidence.

Summary

Remove Chromium from Deno Sandbox to Fix Security

Eliminates a critical escape vector in the Deno sandbox, preventing potential remote code execution and strengthening platform security.

Health Assessment

• • Rapid resolution with minimal changes, indicating efficient review and low complexity.

Summary

Remove git from Deno sandbox allowlist

The change removes the ability for worker scripts to invoke git, closing a critical sandbox escape that could allow authenticated users to run arbitrary OS commands as root. This enhances the security posture of the worker environment.

Health Assessment

• • Fast turnaround with minimal changes, indicating a quick response to a critical security advisory. Single commit, no review rounds, low complexity.

Summary

Add partial index to speed child job listing

Improves job listing queries from seconds to milliseconds, reducing load and improving user experience.

Health Assessment

• • Merged within minutes with no review, minimal changes, low risk.

Summary

Fix bunnative language acceptance in AI flow validation

Ensures AI-generated flow steps using the bunnative script language are no longer rejected by the copilot flow editor, aligning frontend validation with backend language support. This update synchronizes the OpenAPI spec and regenerates schema files, adding a regression test to guard against future drift.

Health Assessment

• • Quick, AI‑assisted fix with minimal scope and no UI changes; the PR was merged within 13 hours, indicating smooth review and low friction.

Summary

Persist per-session preview panel resize width

Ensures the preview panel width set by users persists across page reloads and session switches, improving user experience.

Health Assessment

• • Merged within 11.5 hours with no review back‑and‑forth, indicating a straightforward fix with minimal friction.

Summary

Sync theme into session preview iframes

Ensures session preview iframes reflect live theme toggles, improving visual consistency for users. App iframes remain isolated as intended.

Health Assessment

• • Quick fix with minimal changes, resolved theme sync issue promptly.

Summary

Open test pane when enabling debug

Ensures the debug UI appears immediately when toggling debug mode, improving developer experience by eliminating an extra click.

Health Assessment

• • Single commit, quick fix, minimal scope, no review friction.

Summary

Name draft in AI chat test-run confirmation

Adds dynamic naming to AI chat test‑run confirmation cards, clarifying the target and indicating draft usage, improving user understanding.

Health Assessment

• • Quick AI‑assisted fix merged within 5 hours, minimal code churn, no major review cycles.

Summary

Fix superadmin fork picker for non-member workspaces

Ensures superadmins can view and fork workspaces they are not members of, improving admin workflow and preventing blocked actions.

Health Assessment

• • Quick fix with minimal rework; no blockers and low risk to production.

Summary

Fix autosave for forked script drafts

Forking a script or seeding a new draft from a URL/YAML/JSON import now correctly autosaves and appears in the scripts list.

Health Assessment

• • Quick fix with a clear explanation and test plan, low risk due to small scope and focused change.

Summary

Add datatable migrations for Hub projects

Enables automatic creation and application of database migrations when publishing projects to the Hub, ensuring data consistency and reducing manual schema updates.

Health Assessment

• • The PR involved extensive changes across backend and frontend, with rapid iteration and a single merge, indicating a focused effort to integrate migration support.

Summary

Fix session preview UI elements

Improves user experience by labeling raw app tabs, adding hover grabber, and capping diff drawer size.

Health Assessment

• • Quick turnaround with minimal rework; AI assistance accelerated fixes.

Summary

Add multi-select mode to copilot askUserQuestion

Allows users to select multiple answers in copilot prompts, enhancing flexibility and data accuracy.

Health Assessment

• • Fast 4.5‑hour cycle with single commit and AI assistance, indicating low review friction and minimal risk.

Summary

Scope editors to session workspace

Ensures editor operations target the correct workspace, preventing errors such as "App not found" and ensuring data consistency during AI sessions.

Health Assessment

• • Fast 8‑hour cycle with multiple commits shows efficient review and rapid iteration; large scope across many components but no visible UI changes reduces impact risk.

Summary

Release version 1.753.0

Publishes a new release, updating dependencies, documentation, and API specs, making new features and bug fixes available to users.

Health Assessment

• • Fast automated release with minimal review, indicating efficient release process.

Summary

Add connection sizing guidance to DB health

This PR adds automated guidance for sizing Postgres max_connections based on live worker fleet. It helps operators avoid connection exhaustion and maintain database stability.

Health Assessment

• • The PR was completed quickly with minimal review iterations, indicating a straightforward implementation. The changes span backend logic and frontend UI, but the overall scope remains moderate.

Summary

Clarify workspace fork naming and context

Improves CLI help and defaults for workspace forking, enhancing user experience and reducing errors.

Health Assessment

• • The PR was merged within 1.7 hours with no review iterations, indicating a smooth process. The changes are modest in scope and focus on user‑facing documentation and CLI defaults, posing minimal risk.

Summary

Fix fork error by naming offending item

The PR adds detailed error reporting for workspace forking failures caused by NUL escapes in JSON, enabling users to pinpoint and fix the offending app or flow. This reduces support overhead and improves user experience.

Health Assessment

• • Fast cycle time with a single commit and no review rounds indicates a straightforward bug fix with minimal friction.

Summary

Replicate external secrets on workspace fork

Ensures forked workspaces correctly copy secrets from external backends, preventing access errors and maintaining isolation.

Health Assessment

• • Fast 3.9‑hour cycle, single commit, no review iterations, minimal risk.

Summary

Add condensed top bar for session preview editors

Reduces vertical space in narrow preview panes, improving usability for AI session previews without affecting full-page editors.

Health Assessment

• • Merged within 36 minutes with a single commit, indicating minimal review friction and a straightforward change.

Summary

Preserve worker group tag on Run again

Ensures the worker group tag is retained when re-running jobs, preventing misrouting and improving reliability for users.

Health Assessment

• • Quick AI-assisted fix merged within 5.5 hours, indicating efficient review and low complexity.

Summary

Add scoped preview refresh and multi-target editors

Enhances session preview with scoped refresh, live editors, and pipeline preview, improving performance and user experience.

Health Assessment

• • Rapid iteration with AI assistance delivered a high-impact feature within 5 hours, showing minimal review friction and efficient code integration.

Summary

Fix session preview editor and picker overflow

Improves user experience by ensuring editors and dropdowns fit within the session preview pane, eliminating unwanted page scrolling and visual glitches.

Health Assessment

• • Quick, low‑impact fix with minimal code changes and no review friction.

Summary

Add shared tab system and markdown code blocks

Provides consistent, styled markdown code blocks across the app, unified tab navigation, and improved scrollbars for better readability and interaction.

Health Assessment

• • Rapid iteration with AI assistance led to a quick merge within 3.3 hours, minimal review friction, and a moderate-sized feature addition.

Summary

Add AI chat background jobs tray

Enables background job handling for AI chat, improving responsiveness and providing users with job status, auto‑resume, and approval controls.

Health Assessment

• • Rapid AI‑assisted iteration delivered a large feature set with minimal review friction; extensive unit tests and end‑to‑end verification mitigate risk.

Summary

Bump bundled Go CLIs to clear CVEs

Updates bundled Go CLI binaries to the latest patched versions, eliminating critical CVEs in runtime images.

Health Assessment

• • Rapid security fix with minimal changes and no review needed.

Summary

Refresh picomatch lockfile to fix npm ci

Fixes CI failures by updating the lockfile to include the latest picomatch, ensuring npm ci runs successfully on fresh runners.

Health Assessment

• • Quick lockfile update resolved CI issue with minimal changes and no code modifications.

Summary

Release version 1.752.0

Publishes a new release, updating dependencies, documentation, and configuration files to support the latest features and bug fixes.

Health Assessment

• • Release generated automatically with minimal manual changes, resulting in a quick 12.6‑hour cycle time and low review friction.

Summary

Fix pipeline editor graph and draft issues

Corrects multiple bugs in the pipeline editor, ensuring live edits and deployments reflect accurately, improving reliability and user experience.

Health Assessment

• • Fast turnaround with AI assistance, minimal review, and no blockers.

Summary

Smooth AI chat streaming with typewriter reveal

Enhances chat UX by smoothing bursty AI responses into continuous typing, reducing perceived latency and eliminating freeze‑jump artifacts.

Health Assessment

• • Fast merge with minimal review rounds indicates low friction; the change is medium‑sized but focused on UI, posing low risk.