source-command-fix-ci

To automatically fix CI failures in a monorepo, run the CI check from the repository root to parse lint, type, and test output. The skill groups errors by file, applies inline edits or delegates fixes, and re-runs checks in a verification loop to report fixed versus unfixable issues.

Triage lint and TypeScript errors across multiple packages by running a centralized CI check that categorizes failures deterministically by file. This process sequences fixes to reduce repeat churn and uses delegation rules for complex errors to streamline the correction workflow.

Yes, this CI error fixer works with your bun and turbo monorepo setup by targeting workflows where `bun run ci-check` executes turbo lint, type, and test tasks. It requires command execution directly from the repository root to function correctly.

Automated CI error delegation handles complex test failures by evaluating error counts and complexity, then routing them to a dedicated `ci-fixer` sub-agent. Simpler errors receive inline edits, while the verification loop reruns relevant checks to confirm the fixes.

Repeat churn across different files happens when fixes are unsequenced. The skill mitigates this by deterministically grouping failures by file and error type, sequencing the fixes in batches, and re-running checks to verify results before reporting unfixable issues.

Yes, you can see fixed versus unfixable CI errors because the verification loop re-runs relevant checks after each fix batch. It produces a clear final summary report detailing exactly which lint, type, and test failures were resolved and which require manual intervention.