ground-my-take

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

• Third-party content exposure detected (medium risk: 0.30). Source: the skill reads outsider-authored free text by querying its backlog/queue for items with .slipbox/bin/slipbox evergreen find --status to-discuss, then runs grounding/compass over those retrieved note contents without requiring a trusted first-party selection beyond whatever an outsider may have previously posted into the backlog.