ci-autofix

SUSPICIOUS: the skill is purpose-aligned for CI remediation, but it combines untrusted log ingestion, possible secret exposure to LLM context, execution of repo-controlled scripts, and autonomous commit/push behavior. No malicious installer or third-party credential sink is present, so this is not malware, but it is a medium-risk automation skill that needs strong orchestration guardrails and redaction enforcement.