update

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

• Third-party content exposure detected (high risk: 0.85). In Default/Comprehensive mode Step 2 (“Sync Tasks from External Sources”) and Step 3/Extra Step (“Scan Activity Sources”), the workflow fetches and reads outsider-authored content from external task trackers (e.g., GitHub Issues via gh issue list) and from chat/email/calendar/docs, then ingests the resulting message/task text into the agent for diffing and “Possible Missing Tasks”/memory gap decoding.