Pull Request Explorer

Summary

Add tests for AI key fallback logic

Unit tests validate fallback behavior when a selected AI key cannot serve chat runs, ensuring reliable provider selection.

Health Assessment

• • Fast cycle time and minimal review indicate smooth integration; refactor and tests added with no blockers.

Summary

Agents follow tier changes for newer models

Ensures managed agents automatically use updated model tiers, improving consistency and cost efficiency.

Health Assessment

• • The PR was reviewed quickly and merged within 26 hours, but required several commits to address feedback, indicating moderate iteration but a smooth process.

Summary

Add AI provider selection for chat search and images

Admins can now select which AI provider and image model the chat feature uses for web search and image generation. This gives finer control over costs and performance.

Health Assessment

• • Fast turnaround with minimal review friction; large scope but well‑managed; no breaking changes; high confidence in stability.

Summary

Add tier-aware AI step handling

Ensures AI steps use correct model tiers, preventing stale model usage and improving reliability for users and developers.

Health Assessment

• • The PR was reviewed and merged within 26 hours with a single commit after the initial review, indicating a smooth review process. The change is large in scope but did not trigger multiple review rounds, suggesting manageable complexity.

Summary

Serve published model tiers to chat and agent pickers

Adds an API endpoint to expose model tiers and updates UI pickers to consume them, enabling dynamic tier updates without redeploy.

Health Assessment

• • PR introduced a new API endpoint and updated UI components to consume dynamic model tiers, with a single review cycle and moderate scope, indicating a well‑managed feature addition.

Summary

Add bot-only channel filter to Slack dropdowns

Users now see only relevant Slack channels, improving UX and reducing API load.

Health Assessment

• • Fast review and merge with minimal rework; well-tested changes keep risk low.

Summary

Add service account authentication to Google Slides piece

Enables Google Slides integration to use service account JSON keys, expanding authentication options and improving reliability for users.

Health Assessment

• • Long cycle time and multiple commits suggest significant rework; the added authentication flow may affect existing integrations and requires careful testing.

Summary

Backfill action classification for 226 pieces

Adds classification tags to actions and triggers across 226 pieces, improving metadata for the piece selector and enabling better categorization.

Health Assessment

• • The PR added over 2,500 lines across 20 files, with a quick review and merge, indicating efficient collaboration. The large scope suggests significant impact on the piece catalog metadata.

Summary

Fix agent steps to use only available models

Ensures AI agent steps only use models supported by Activepieces AI credits, providing clear failure messages when unsupported models are selected, preventing internal errors and improving user experience.

Health Assessment

• • Rapid iteration with three commits in under an hour, minimal review friction, and a small scope indicates a straightforward bug fix with low risk.

Summary

Fix chat taint guard and memory save behavior

Improves security by ensuring data reads are properly tainted across turns and memory updates are confirmed, preventing prompt injection and unintended memory persistence.

Health Assessment

• • Rapid resolution with minimal iterations; quick review and merge within 1.1 hours, indicating high confidence in the fix.

Summary

Fix legacy CommonJS code steps in Deno

Legacy CommonJS code steps now run correctly in Deno, preventing production failures during migration. This maintains high success rates for flows during the Deno rollout.

Health Assessment

• • Quick, single-commit fix with minimal review time and no breaking changes, indicating a low-risk, high-impact patch.

Summary

Fix Docker prune to remove core devDependencies

The change removes core-piece devDependencies from the Docker build, preventing bun install failures. This stabilizes staging image builds and restores CI reliability.

Health Assessment

• • Quick fix with minimal changes, resolved CI failures promptly.

Summary

Fix cross‑conversation file read vulnerability

Prevents the code tool from accessing files belonging to other conversations, eliminating a potential data exposure risk.

Health Assessment

• • Quick 1.2‑hour turnaround with minimal changes, low review friction, addressing a critical security issue.

Summary

Add AI actions and output schemas for Jotform

Adds 44 agent-facing actions and output schemas to the Jotform integration, enabling AI-driven automation and improved data handling.

Health Assessment

• • Rapid review and merge within 1 hour, indicating high confidence and minimal friction.

Summary

Add delete confirmation gate for chat actions

Prevents accidental or malicious deletion of data via chat, enhancing data integrity and security.

Health Assessment

• • Fast turnaround with minimal rework; quick fix to a critical security issue.

Summary

Fix Notion piece lint errors

Resolved linting issues in the Notion integration, ensuring code quality without altering functionality.

Health Assessment

• • Quick, single‑commit fix with no functional impact; minimal review and merge time indicates low risk and high confidence in stability.

Summary

Fix chat credit exhaustion handling

Ensures chat stops when credits are depleted, preventing overuse and billing errors for users.

Health Assessment

• • Rapid iteration with multiple commits in under 2 hours indicates focused development and quick resolution of credit logic bugs.

Summary

Add dynamic AI model tier resolution

Enables agents and chat turns to resolve AI model tiers from a published file, allowing runtime updates without redeploys.

Health Assessment

• • Fast review and single iteration indicate a smooth process; moderate scope but no breaking changes.

Summary

Deprecate Straico piece due to service shutdown

Mark the Straico integration as deprecated to hide it from new users while preserving existing flows.

Health Assessment

• • Rapid merge with minimal changes and no review required.

Summary

Add AI actions and output schemas for Tavily piece

Introduces agent‑facing actions for crawling, mapping, research, and usage, with comprehensive output schemas, enabling AI workflows without new authentication or scopes.

Health Assessment

• • Fast cycle time (0.3h) and immediate review (0.1h) indicate minimal friction; the PR added 626 lines across 12 files, a large but straightforward change with no breaking or security impact.

Summary

Add AI actions and output schemas to NocoDB piece

Adds 12 new AI-facing actions to the NocoDB integration, enabling automated workflows and richer data interactions for users.

Health Assessment

• • Rapid review and minimal rework indicate high confidence in the implementation; the large addition of actions expands automation capabilities for users.

Summary

Fix mislabeling of Reoon Verifier actions in tool picker

Corrects action classifications to reflect read/write behavior, improving agent tool picker accuracy and ensuring proper permissions.

Health Assessment

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

Summary

Add AI actions and output schemas to APITemplate.io

Adds new AI‑facing actions and output schemas for APITemplate.io, enabling richer PDF and template manipulation for users. This expands the piece’s capabilities without requiring new scopes.

Health Assessment

• • PR merged quickly with minimal review, indicating straightforward implementation.
• • No significant rework or blockers observed.

Summary

Add AI actions and output schemas to Cal.com piece

Adds 25 agent-facing actions for Cal.com integration, enabling AI agents to manage events, bookings, schedules, and more, expanding functionality for users.

Health Assessment

• • The PR was reviewed and merged within 44.5 hours with only one round of changes, indicating efficient collaboration.
• • Despite the large scope, the quick turnaround suggests the changes were straightforward and well-tested.

Summary

Revive Instagram Business piece with Graph v23.0

Enables Instagram Business integration with updated Graph API, adds 27 verified actions, improves reliability, and requires re-consent for new OAuth scopes, enhancing platform functionality for users.

Health Assessment

• • The PR had a long cycle time and required multiple commits after the initial review, indicating significant rework. The large code addition and breaking changes increase risk for downstream consumers.

Summary

Add AI-ready Dropbox actions for files and batches

This PR expands the Dropbox integration with 26 new AI-focused actions, giving agents richer file handling capabilities and simplifying batch operations.

Health Assessment

• • The PR received a quick review but had a long cycle time due to extensive changes; the large scope and many new actions increase risk of regressions.

Summary

Add 59 Asana AI agent actions

Expands Asana integration with 59 new AI-driven actions, enabling advanced workflow automation and paid‑plan features for users and teams.

Health Assessment

• • PR required multiple commits to address feedback, indicating moderate complexity, but the initial review was rapid. The large number of added actions represents a significant feature expansion.

Summary

Add 65 AI agent actions for Asana tasks, projects, sections and tags

Introduces a comprehensive set of AI‑driven automation actions for Asana, enabling users to create, update, and manage tasks, projects, and related entities directly through the platform.

Health Assessment

• • The PR added a large amount of new code (≈4,900 lines) and was merged after a long cycle time of 90 hours, indicating potential delays in initial planning or testing. The review was quick, suggesting the changes were straightforward to understand once presented. Overall, the feature introduces significant new functionality with a high risk due to its size and extended review cycle.

Summary

Add Mastodon triggers and enhanced Post Status

Enables flow‑builder users to trigger on Mastodon events and post richer statuses, improving automation and user experience.

Health Assessment

• • Large scope with 20 files and 681 added lines, but quick review and no blockers suggest a well‑managed change.

Summary

Add image editing in chat conversations

Enables users to edit existing images within chat, enhancing interaction and reducing repetitive image generation.

Health Assessment

• • The PR was merged within 2.2 hours with only a single review, indicating a smooth, low-friction process. The change touches many files but required minimal rework, suggesting a well-scoped feature with moderate risk.

Summary

Add model tier reader from published file

Enables dynamic AI tier configuration via CDN, reducing deployment overhead and allowing model updates without redeploying the service.

Health Assessment

• • Fast cycle time and minimal review comments indicate smooth integration; single review round and quick merge suggest low friction.

Summary

Add 19 WhatsApp actions and 4 triggers

Enables comprehensive WhatsApp integration with 22 actions and 4 webhook triggers, improving automation and user engagement.

Health Assessment

• • The PR required many commits and a long cycle time, indicating substantial effort and potential risk, but the first review was rapid, suggesting initial acceptance of the changes.

Summary

Add AI-agent atomics to pipedrive and mailer-lite

Introduces 69 AI‑driven atomics across two integration pieces, expanding granular AI actions while preserving existing flow behavior.

Health Assessment

• • PR had a long cycle time but a rapid first review, suggesting extensive testing and iterative commits to refine the feature.

Summary

Add Microsoft Outlook agent atomics for mail operations

Adds 45 AI‑centric actions enabling agents to read, send, organize, and manage Outlook mail, improving automation and reducing manual steps.

Health Assessment

• • PR introduced extensive new functionality with 45 atomics, required multiple quick fixes, but received rapid review and merged after ~5.6 days.

Summary

Fix Destructive action badge styling

Aligns the destructive action badge with other action badges to improve UI consistency and reduce confusion for users.

Health Assessment

• • Rapid resolution with minimal code changes and a quick review cycle indicates low complexity and high confidence in the fix.

Summary

MCP page uses platform server, Pieces merged into Tools

Users can now access MCP features via the platform server, simplifying navigation and consolidating tools, improving usability and reducing confusion.

Health Assessment

• • The PR required extensive iteration with many commits after review, indicating significant rework, but the review was quick and the final merge was straightforward.

Summary

Fix MCP tool switch binding across servers

Ensures that disabling tools in the platform or project scope applies globally across all MCP servers, preventing unintended tool execution and billing.

Health Assessment

• • The PR required multiple iterations after a delayed review, indicating complex changes to MCP server logic and extensive testing.

Summary

Enable project members to use MCP server gated on READ_MCP

Adds fine‑grained permission checks so that only users with READ_MCP can access the MCP server, tightening security and clarifying role boundaries for project members.

Health Assessment

• • The PR received a quick initial review but required several follow‑up commits to resolve security and permission logic, indicating significant rework. The large code churn and security‑sensitive changes elevate overall risk despite the short review time.

Summary

Generate images with platform AI provider

Adds image generation capability using existing AI provider without requiring a separate key, improving user experience and simplifying setup.

Health Assessment

• • Fast cycle time (1.6h) and immediate review indicate a smooth process; the PR added significant new functionality with moderate scope and no major blockers.

Summary

Retire old CI workflow, promote CI v2

Streamlines CI pipeline by consolidating workflows, improving reliability and reducing maintenance overhead.

Health Assessment

• • Fast review and merge cycle indicates smooth process; minimal rework; consolidated CI reduces maintenance.

Summary

Ensure project role names unique per platform

Prevents accidental deletion of roles and preserves user assignments by enforcing unique role names per platform, eliminating silent data loss.

Health Assessment

• • The PR required 5 days to merge after a rapid review, reflecting extensive testing and a large scope of changes across backend, frontend, and database migrations.

Summary

Add web search across all AI providers

Enables chat agents to search the web without requiring a Tavily key, improving user experience and reducing setup overhead.

Health Assessment

• • Fast cycle time and minimal review indicate a straightforward change.

Summary

Rebuild role dialog with view/edit permissions

Enhances role management UI by replacing old permission controls with view/edit checkboxes, adding a people tab, and improving permission visibility. Enables admins to configure roles more accurately without breaking existing APIs.

Health Assessment

• • Large UI overhaul with extensive changes, but quick review and minimal back‑and‑forth; long cycle time likely due to scheduling or other priorities.

Summary

Rename Roles tab and rebuild roles card

This PR updates the Roles & Access UI, renaming the tab to "Roles" and redesigning the roles card to improve clarity and usability for admins, enabling better role management and error handling.

Health Assessment

• • The PR had a long cycle time due to integration with main, but the review was quick and the changes were straightforward UI refactor with no API impact.