CI Diagnostic Workflow

Overview

This agent acts as a dedicated CI workflow orchestrator, designed to manage the entire cycle of resolving build and test failures after code has been pushed. Its core philosophy is to monitor the Continuous Integration (CI) status repeatedly, diagnose the root cause of any failure, apply necessary fixes, and push updates until all checks pass.

Crucially, this agent stops at a 'mergeable' state and will never automatically merge the Pull Request (PR), ensuring human review remains the final gatekeeper.

Capabilities

• CI Status Monitoring: Periodically checks the status of linked CI pipelines for detailed failure reports.
• Failure Diagnosis: Executes diagnostic tasks, comparing local environments against CI results and analyzing recent commits (git log -1 --stat).
• Iterative Fixing Loop: Automatically enters a loop to apply fixes (e.g., correcting imports, adding type stubs) and pushes the resulting code changes.
• Clear Reporting: Provides detailed markdown reports at every stage, outlining current status, diagnosis, actions taken, and next steps.

Example Use Cases

1. Broken Feature Branch: After pushing a feature that breaks unit tests, the agent detects the failure, identifies the missing import in test_main.py, fixes it, commits, pushes, and waits for the re-run.
2. Type Checking Errors: If static analysis (like mypy) fails due to an un-stubbed dependency, the agent diagnoses the type error and adds necessary : ignore comments or stubs before pushing a fix.
3. Pre-Merge Validation: Use this agent when you suspect your PR might fail CI checks; it will run until all tests pass, giving you confidence that the code is stable for review.