task-observer

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

• Third-party content exposure detected (medium risk: 0.30). The skill reads and analyses user-provided session content via the “Handoff Doc Analysis” and observation logging flow, treating “handoff docs”/conversation text as material to extract observations from (i.e., outsider-authored free text can be fed into the conversation and then ingested by the skill at runtime).