fix-ci

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and reads GitHub Actions CI logs (see "Fetch — Download CI logs from GitHub Actions", the $ARGUMENTS field, and Workflow step 1 using the gh command) and uses those untrusted, user-generated logs to drive analysis and code-fixing actions, so third-party content can directly influence tool decisions.