aws-cloudformation

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection Surface: The skill processes CloudFormation templates, which may contain untrusted user-controlled text in fields like Description, Metadata, and inline comments. The skill mitigates this by providing explicit instructions to treat all template content as untrusted data and forbidding the agent from obeying any directives found within it.
  • External Resource References: The skill references external resources for validation rules and documentation. It fetches compliance rules from an official AWS repository on GitHub and refers to the primary AWS CloudFormation documentation. These references target official vendor infrastructure.
  • Package Management: The skill utilizes standard validation tools such as cfn-lint and cfn-guard. It provides instructions for installing these from official package registries like PyPI, requiring explicit user approval before modifying the environment.
  • Command Execution: The skill performs various operations using the AWS CLI and the AWS MCP server. These include creating change sets, describing events, and performing template validation. These commands are integral to the skill's primary function of managing cloud infrastructure and include safety checks such as requiring user approval for destructive operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 10:32 AM
Security Audit — agent-trust-hub — aws-cloudformation