deep-research

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

• Third-party content exposure detected (low risk: 0.10). The skill’s stage-2 (scripts/iterate.py) only prints a fetch list and records host-returned evidence_round_*.json (URLs/excerpts/relevance) into plan.json, while the actual outsider-authored text ingestion happens in the host web-fetch tools rather than within this skill’s required runtime workflow.