agent-autonomy-kit-hardened

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly directs agents to read and act on user-generated third-party channels (e.g., "Team Coordination — Agents communicate through Discord" and the configured team channel in README/Setup) and to perform external research tasks (e.g., "Research competitor X pricing"), meaning the agent will ingest untrusted social/web content as part of its heartbeat/workflow and that content can influence tool use and decisions.