weather-hardened

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 the agent to fetch and parse live data from public services (wttr.in and Open‑Meteo) and SAFETY.md documents workflows that ingest third‑party geocoding JSON (potentially user‑supplied endpoints), meaning the agent will read and act on untrusted web content that could carry indirect prompt‑injection payloads.