The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts and Python code to execute AWS CLI commands and utilize the boto3 library for resource enumeration and configuration auditing. These operations are intended for auditing purposes and align with the stated goal of compliance checking.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources (AWS APIs), including IAM credential reports and resource configuration metadata. This introduces a surface for indirect prompt injection if resource names or report contents are controlled by an adversary. However, the skill employs structured data processing and includes safety warnings regarding the handling of evidence.
Ingestion points: references/detailed-guide.md (via aws iam, aws ec2, and boto3 calls).
Boundary markers: Present in SKILL.md as guidance for scope verification and handling sensitive material.
Capability inventory: Execute AWS CLI commands, run Bash scripts (using jq, awk), and write local JSON reports in compliance-report.py.
Sanitization: Data is handled through structured interfaces like boto3 and jq, reducing the risk of accidental command injection during parsing.

aws-compliance-checker