fetch-ci-build

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow and scripts (SKILL.md Step 3–4 and scripts/fetch_buildkite_failures.py's fetch_raw_log and scripts/fetch_circleci_failures.py's fetch_job_log) explicitly fetch and parse CI build logs and job output from third-party CI provider APIs (Buildkite/CircleCI/GitHub) which are untrusted, user-generated content that the agent reads and uses to drive diagnostics and fix actions.