automate-me

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

• Third-party content exposure detected (high risk: 0.70). The required runtime workflow includes “Mine their history” by having subagents read transcripts from a workspace-scoped path provided by the host, which can include outsider-authored free-form chat/transcript text (e.g., other participants’ messages) that the LLM then ingests for pattern mining.