trigger-realtime

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

• Third-party content exposure detected (high risk: 0.70). The workflow uses Trigger.dev realtime stream data (aiStream parts from useRealtimeStream / aiStream.pipe(completion)), where the streamed text is ultimately derived from runtime model/task output (user-controlled payload.prompt), so arbitrary user/outsider free text can be ingested into the agent’s LLM context via streamed content.