security-report-writer

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime report generation ingests the pentest “raw notes”/finding content (e.g., from issues/PRs/discussion threads or other external test artifacts) into the LLM context to “turn raw notes into a professional internal pentest report,” which can include outsider-authored free text; the SKILL.md itself doesn’t show a strict exclusion of such inputs.