adk-deploy-guide

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

• Third-party content exposure detected (high risk: 1.00). The skill's event-driven reference (references/event-driven.md) shows runtime endpoints that decode Pub/Sub, Eventarc, and BigQuery trigger payloads and pass the resulting text directly into the agent via _run_agent (e.g., base64-decoded Pub/Sub data -> _run_agent), so untrusted external messages can be ingested and influence agent actions.