openarc-patrol

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

• Third-party content exposure detected (high risk: 0.90). Outsider free text enters the LLM context via the required runtime step where the agent “judges” each batch item by applying judge_prompt to the post’s text (from the openarc pending --limit 15 JSON), and that post text is explicitly treated as untrusted.