lead-researcher

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

• Third-party content exposure detected (high risk: 0.70). The required runtime workflow reads outsider-authored free text from leads.csv (and potentially icp.json) via scripts/lead_qualifier.py using Path(args.leads).open(...) and csv.DictReader, then prints/serializes fields like reason and matched_signals into the agent’s output context.