autofix-bot-api

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

• Third-party content exposure detected (high risk: 0.80). The skill fetches analysis results and unified-diff "fix_result.patch" from the external Autofix Bot API (see SKILL.md Step 5 and the GET /analysis/{id} responses, and scripts/poll_analysis.sh), and explicitly instructs applying those patches (echo "$FIX_PATCH" | git apply), so third-party API-generated content is read and can directly change agent actions.