cloud-architect

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides architectural templates and command-line snippets for cloud management. All provided code (Terraform, Python, HCL, and Bash) follows standard industry practices and does not include any malicious payloads.
  • [COMMAND_EXECUTION]: The skill contains numerous templates for AWS CLI (aws ec2), Azure CLI (az network), and Terraform. These are intended for cloud administration and are used within a descriptive workflow for architectural design and migration verification.
  • [DATA_EXFILTRATION]: The skill explicitly instructs against storing credentials in code or public repositories. It correctly identifies the use of environment variables and IAM roles for secret management. Access to the AWS metadata service (169.254.169.254) in references/cost.md is a standard practice for handling spot instance interruptions.
  • [PROMPT_INJECTION]: No patterns associated with prompt injection, role-play bypasses, or instruction overrides were found. The workflows are structured to provide professional architectural advice.
  • [EXTERNAL_DOWNLOADS]: The skill references standard cloud management tools (Terraform, Pulumi, rclone) and official documentation. No suspicious third-party scripts or unverified remote payloads are included.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 10:31 AM
Security Audit — agent-trust-hub — cloud-architect