book-publishing

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

• Third-party content exposure detected (medium risk: 0.30). The skill reads outsider-authored free text from the user prompt (e.g., topic/genre/publishing goal/distribution platforms and attached manuscript content) during orchestration “Extract from user input” in Phase 1 and then feeds derived details to the agents via workspace files and SendMessage.