peer-impact

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

• Third-party content exposure detected (low risk: 0.10). In scripts/peer_impact.py, the runtime LLM input is limited to JSON read from the user-supplied --candidates file (open(args.candidates) + json.load), whose fields like evidence, note, name, and rejected_because are rendered to output; there is no external fetching/search or ingestion of outsider free text beyond that file.