modelopt-run-agentx

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

• Third-party content exposure detected (low risk: 0.10). The required workflow fetches public AgentX trace and prompt-reconstruction assets from a Hugging Face dataset via the harness loader, and those outsider-authored traces are replayed to the server rather than read by the agent, so this is a selected public artifact the agent must actively fetch (low).