learn

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

• Third-party content exposure detected (high risk: 0.80). The skill's "Reading" mode in the Process/Deliver Content section explicitly tells the agent to "curate relevant articles, documentation, or book chapters (use web search for up-to-date content)," which requires fetching and interpreting untrusted public web content as part of its workflow and could let external instructions influence its guidance.