packet-capture-analysis

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs making external lookups (RDAP / reverse DNS and similar public queries) to label IPs (see "IP を意味付けする", "どこまで外部照会するか", and the recommended script "label_endpoints.py"), and those third‑party responses are read and used to assign labels that materially affect triage decisions and reports.