The Agent Skills Directory

Command Execution: The skill utilizes the AWS CLI (aws sts, aws iam, aws s3api, aws securityagent) to manage cloud resources. This is necessary for the skill's primary function of infrastructure provisioning.
Privilege Escalation Consideration: To function correctly, the skill creates IAM service roles and attaches policies. It follows security best practices by including an aws:SourceAccount condition in the trust policy to mitigate confused-deputy risks and by applying a strict public access block to created S3 buckets.
Indirect Prompt Injection Surface: The skill ingests data from local configuration files (.security-agent/config.json) and external AWS service responses (e.g., list of agent spaces). While this provides a potential surface for indirect prompt injection, the risk is mitigated by the skill's specific focus on infrastructure management within the user's controlled environment.
Ingestion points: Reads from .security-agent/config.json and parses outputs from aws securityagent list-agent-spaces.
Boundary markers: None identified.
Capability inventory: Performs IAM role creation, S3 bucket provisioning, and local file writes.
Sanitization: Standard AWS CLI output parsing is used; however, no explicit validation of names returned from the service is described.
File System Interaction: The skill maintains local state in a .security-agent/ directory. It automatically generates a .gitignore file to ensure that local configuration remains private and is not accidentally committed to version control.

setup-security-agent