monitor-personas

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests public, user-generated content (YouTube, RSS, arXiv, social feeds/dogpile and other discover-* skills) — see SKILL.md and curate.py (calls to _run_skill for ingest-youtube, discover-*, social-bridge, dogpile) and monitor_commands/backfill_taxonomy (list_channel_video_ids, classify()) — and that content is parsed and used to drive classification, memory learning, taxonomy stamping, training and autonomous pipeline actions, so untrusted third‑party content can materially influence tool use and decisions.