academic-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 mandated workflow (SKILL.md and example.md) explicitly requires using web_search (count=20) and web_fetch on primary sources to ingest open/public web content (e.g., news sites, forums, vendor docs), meaning the agent will read and act on untrusted third‑party material as part of its decision-making.