chat-failure-audit

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow ingests the “transcript/log you provide” (Step 0: transcript can be a URL fetched at runtime or pasted session turns), and those turns may include outsider-authored free text, which the agent then reads and quotes into the LLM context during Steps 1–4.