teach

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

• Third-party content exposure detected (medium risk: 0.30). In topic mode, outsider-authored free text can be ingested indirectly via user-provided RESOURCES.md entries/URLs because the skill later “fetch[es] and confirm[s] URLs before adding” and “re-fetch[es] the cited source” for verification/staleness, so a user can supply attacker-controlled pages that the runtime will read.