security-runner

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

• Third-party content exposure detected (medium risk: 0.65). At runtime the skill calls the GitHub API for Dependabot alerts (gh api repos/"$REPO"/dependabot/alerts ... --jq ...), and the resulting JSON fields like security_advisory.summary and html_url are then interpolated into the LLM-readable report text (DETAIL_LOG and REPORT_PATH) and optionally posted via curl—this content is authored by third parties (Dependabot/GitHub security advisories), not by the operating user.