develop

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

• Third-party content exposure detected (high risk: 0.70). The skill's Phase 8 "Document (Blog Post)" explicitly tells the agent to "Use reddit to find related community discussions," which directs the agent to fetch and interpret untrusted, user-generated third-party content as part of its workflow and could materially influence writing/decisions.