agent-platform-alert-configuration

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow runs scripts/analyze_traffic.py in “live” mode, which queries Cloud Monitoring time-series metric values and then feeds the resulting numeric stream into compute_metrics_and_classify(); while it ingests outsider-authored content only indirectly via metrics, that content comes from runtime traffic rather than trusted first-party docs.