resonate-human-in-the-loop-pattern-python

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

• Third-party content exposure detected (low risk: 0.10). The workflow described uses ctx.promise() to suspend and is resumed only when the external resolver calls r.promises.resolve/reject (with webhook/FastAPI route examples that parse JSON fields), so outsider-authored free text is only ingested as explicit decision payload data provided by that caller rather than by the required workflow spontaneously reading arbitrary attacker text.