research-ideation

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (SKILL.md Step 2 "Build a Literature Tree" and linked references/literature-tree.md and paper-reading.md) explicitly instructs the agent/user to "collect papers", "search for papers", "follow citation chains" and "as you read papers, extract technical challenges" — i.e., ingest and interpret public third‑party papers/web content that can directly influence problem selection and next actions.