scientific-slides

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

• Third-party content exposure detected (high risk: 0.80). The skill repeatedly instructs the agent to use the "research-lookup" capability to search for and ingest public literature/papers (see "Research and Literature Review (Use research-lookup skill)" and numerous "use research-lookup" mentions), so the agent will fetch and read external third‑party content as part of its workflow, exposing it to potential indirect prompt injection.