Pull Request Explorer

Summary

Restore WhatsApp contact info requests for BSUID

Enables agents to request phone numbers again for eligible WhatsApp conversations, improving data completeness and reducing manual steps.

Health Assessment

• • Long review cycle with multiple merge commits indicates integration complexity and potential coordination overhead. Automated AI review contributed to thorough checks but extended the overall cycle time.

Summary

Install Captain toolsets from GitHub

Adds API and UI to preview and install Captain toolsets from GitHub, enabling admins to manage toolsets directly from the dashboard and improving extensibility.

Health Assessment

• • Rapid review with many automated comments led to extensive iteration; large code churn and multiple review rounds indicate high complexity and risk, though the review process was efficient.

Summary

Add toolset manifest install service

Adds backend support for installing Captain toolsets from the chatwoot/tools repository, enabling automated toolset import and management for assistants.

Health Assessment

• • The PR underwent extensive automated review and multiple iterations, indicating complex logic and careful validation.
• • The large number of added lines and repeated review comments suggest a high level of scrutiny and potential integration risk.

Summary

Add static headers to custom tools

Enables custom tools to send static HTTP headers and support additional HTTP methods, improving integration flexibility for external APIs.

Health Assessment

• • The PR had a slow overall cycle time but a quick initial review; however, it required many subsequent commits and extensive automated review comments, indicating significant iteration and complexity.

Summary

Fix deadlock by dropping FK on campaign_recipients

Removes a problematic foreign key that caused deadlocks in busy installations, improving database reliability and uptime for users.

Health Assessment

• • The PR was reviewed and merged within 11 hours, with minimal code changes and no new tests, indicating a straightforward, low‑risk fix.

Summary

Add monitor templates to reports

Administrators can start conversation monitors from built‑in templates, streamlining setup and improving monitoring efficiency.

Health Assessment

• • The PR underwent multiple review cycles and required many commits after the first review, indicating significant rework. The large code change and slow cycle time suggest high integration risk.

Summary

Add Captain conditions to automation rules

Enables admins to create automation rules based on semantic message detection, improving rule accuracy and reducing manual keyword configuration.

Health Assessment

• • The PR required extensive review and multiple iterations, indicating complex changes and potential integration challenges.

Summary

Apply macro to multiple conversations at once

Agents can now run a macro on multiple conversations simultaneously, improving efficiency during outages or maintenance windows.

Health Assessment

• • Review took over 69 hours with two commits after the first review, indicating moderate complexity but a quick final approval.

Summary

Add shared client for Jev tracing in Langfuse

Enables consistent telemetry for Jev calls, improving observability and cost tracking.

Health Assessment

• • The PR required several commits after the initial review to align Jev spans with OpenTelemetry conventions, indicating moderate complexity. Review time exceeded 67 hours, but final approval was achieved after addressing telemetry and error handling concerns.

Summary

Update rack-proxy to patched version for security

Fixes a security advisory by bumping rack-proxy to a patched release, ensuring CI security scans pass and preventing potential vulnerabilities in the development environment.

Health Assessment

• • Quick, single-file change with minimal review and no production impact.

Summary

Add WhatsApp campaign history, batch sending, analytics

Enables agents to view WhatsApp campaign history in conversations, processes large campaigns in batches, and provides analytics for paid plans, improving agent efficiency and customer engagement.

Health Assessment

• • The PR had a long cycle time of 11 days but was reviewed quickly by an automated AI tool, indicating minimal human review friction.
• • The large scope and numerous merges suggest significant rework, raising risk for integration issues.