feynman

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

• Third-party content exposure detected (high risk: 0.90). The skill's "With reference material" workflow explicitly asks the user to "Paste or attach the source — article, paper, docs, notes, transcript" and then instructs the agent to "audit — using the source as ground truth," meaning it ingests and acts on arbitrary/untrusted third-party content supplied at runtime.