The Agent Skills Directory

[SAFE]: The skill consists of informational documentation, AWS IAM policy templates, and administrative Python scripts. The content is aligned with AWS security best practices such as the Well-Architected Framework and CIS Benchmarks.
[COMMAND_EXECUTION]: The scripts scripts/agent.py and scripts/process.py utilize the standard AWS SDK (boto3) to perform IAM auditing and management tasks. These operations are restricted to standard AWS API calls for role enumeration and policy attachment, which are intended functions for the skill's purpose.
[EXTERNAL_DOWNLOADS]: The skill references official AWS documentation and trusted GitHub repositories (e.g., aws-samples) for reference purposes. No external code is downloaded or executed at runtime.
[DATA_EXFILTRATION]: The auditing scripts generate local JSON reports containing IAM metadata (role names, ARNs, and boundary status). There is no evidence of sensitive data being exfiltrated to external or untrusted domains.
[PRIVILEGE_ESCALATION]: The skill is specifically designed to mitigate privilege escalation by providing templates and tools to enforce permission boundaries that cap maximum permissions for IAM entities.

implementing-aws-iam-permission-boundaries