ccmd

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 caching the stdout of commands that fetch network resources (e.g., "When performing repeated network requests (e.g., curl" and the example x ccmd 1h -- curl https://api.example.com/data), so the agent will fetch and consume arbitrary third‑party web content (untrusted) as part of its workflow, which could materially influence subsequent actions.