launching-evals

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

• Third-party content exposure detected (high risk: 0.70). The skill's required workflow explicitly instructs the agent to fetch and read external, public resources — e.g., "Read the model's HuggingFace model card" (references/analyze-results.md) and to fetch troubleshooting docs from raw.githubusercontent.com via WebFetch (references/debug-failed-runs.md) — which are untrusted, user/third‑party web content and are used to influence validation, debugging, and follow-up actions.