deep-research

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md Methodology step 2 ("Web research: Use available web tools to find current best practices and documentation") explicitly requires fetching and interpreting open web content, exposing the agent to untrusted third‑party pages whose instructions could materially influence its analysis and actions.