automate-me

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

• Third-party content exposure detected (high risk: 0.70). The runtime flow explicitly “mines their history” by reading agent-transcripts/ from the active workspace and feeding mined patterns/evidence into the agent’s LLM context; those transcripts can include outsider-authored free text (e.g., other participants’ messages in chats/issues/PR discussions) that the user did not choose to introduce.