github-ci-fix

This skill is broadly aligned with GitHub CI/PR repair and uses expected GitHub tooling, so it does not look malicious. The main risk is operational: it combines untrusted GitHub content ingestion with autonomous code changes, pushes, and PR actions, and it also chains into another skill not shown here. Overall classification: SUSPICIOUS due to autonomy and indirect prompt-injection exposure, not due to credential theft or overt exfiltration.