creating-production-vpc-multi-az
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [AWS Command Execution]: The skill uses the
call_awstool to execute a sequence of AWS CLI commands for provisioning infrastructure. This includes creating VPCs, subnets, NAT gateways, and route tables. This is the primary function of the skill and aligns with AWS infrastructure management practices. - [Identity and Access Management]: The procedure includes steps to create an IAM role and attach an inline policy using
aws iam create-roleandaws iam put-role-policy. This is specifically used to grant VPC Flow Logs permission to write to CloudWatch Logs, which is a standard requirement for network monitoring. - [Security Group Configuration]: The skill creates security groups and populates ingress rules based on user-provided CIDR blocks (
allowed_web_cidrsandssh_allowed_cidrs). It includes a built-in safety check that warns the user if they attempt to open web access to the entire internet (0.0.0.0/0), promoting secure-by-default configurations. - [Indirect Prompt Injection Surface]: The skill ingests user-defined parameters such as
vpc_nameandallowed_web_cidrswhich are interpolated into AWS CLI commands. While this is necessary for dynamic infrastructure creation, it represents a standard input surface where user-provided data influences the final commands executed via the agent's context.
Audit Metadata