skill-security-checklist

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow runs assets/generate-audit-report.sh, which executes multiple bash detectors (e.g., detect-stack.sh, detect-architecture.sh, detect-infrastructure.sh, analyze-complexity.sh, detect-security-risks.sh, assess-tech-debt.sh) that read and emit free-form text from the target project’s files (including arbitrary source code/comments/config like README.md, .env*, and other repo text) into the LLM context as YAML code blocks; since that project content is not authored by the operating user, it is outsider-authored free text.