google-app-engine

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): The content consists of instructional text and templates without any patterns suggesting attempts to override agent behavior or bypass safety filters.
  • [Credentials/Data Exposure] (SAFE): No sensitive credentials or secrets are hardcoded. The skill uses descriptive placeholders and explicitly instructs users to use Google Cloud Secret Manager instead of environment variables for secrets.
  • [External Downloads/RCE] (SAFE): All suggested dependencies are standard, well-known libraries from PyPI (e.g., gunicorn, google-cloud-secret-manager, google-cloud-storage). No remote script execution patterns (like curl | bash) were detected.
  • [Privilege Escalation] (SAFE): IAM policy commands in the troubleshooting guide are standard for the Google App Engine environment and are intended for the user to execute to resolve legitimate permission errors.
  • [Indirect Prompt Injection] (LOW): 1. Ingestion points: app.yaml handlers and static file configuration paths. 2. Boundary markers: Absent. 3. Capability inventory: Instruction-based gcloud CLI commands in documentation. 4. Sanitization: Not applicable to this documentation-centric skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:46 PM