adk-observability-guide

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The docs explicitly mandate always-on tracing and optional prompt-response logging that uploads full agent interactions (including multimodal content) to GCS/BigQuery/Cloud Logging with long retention, which constitutes deliberate centralized capture/exfiltration of user data and PII if enabled in production.