book-reader-conversations

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

• Third-party content exposure detected (high risk: 0.75). The workflow is designed to run “reader conversations” with recruited outsiders, and the runtime conversation transcripts/notes (free-form messages from those readers) would be ingested into the agent’s LLM context for synthesis (Step 4–5), creating an indirect prompt-injection surface from outsider-authored text.