distill-skill

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text from the recorded CC session (e.g., other participants’ messages) is appended into trace.jsonl as friction/user_intent and then fed to the LLM via the required /edit-tool handoff using draft.yaml and anti-patterns.md (runtime prompt construction includes that content).