always-verify-gcp

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is designed with a strong security posture, implementing a 'Trust but Verify' workflow for Google Cloud Platform operations. It includes mandatory steps for documentation lookup, parameter verification, and user approval, minimizing the risk of accidental or malicious command execution.
  • [COMMAND_EXECUTION]: The skill correctly uses run_shell_command to interface with official GCP CLI tools (gcloud, bq, gsutil). It includes safety precautions such as checking for --dry-run support and presenting command impact to the user before final execution, adhering to best practices for cloud resource management.
  • [PROMPT_INJECTION]: While the skill processes user instructions and external documentation (via the search_documents tool), the risk of indirect prompt injection is effectively mitigated by the structured verification workflow and the requirement for explicit human-in-the-loop confirmation before any state-changing command is executed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 06:49 AM