The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/agent.py script executes system utilities such as nmap, systemctl, and package managers (apt, yum) to automate compliance auditing. These calls are constructed using argument lists via subprocess.run, which follows secure practices for preventing shell injection.
[METADATA_POISONING]: A minor discrepancy exists between the author listed in the SKILL.md frontmatter (mahipal) and the copyright holder in the LICENSE file (mukul975). This appears to be a common documentation error rather than a malicious deception.
[INDIRECT_PROMPT_INJECTION]: The skill has a defined surface for ingesting external data for compliance assessment.
Ingestion points: The audit_compliance_status function in scripts/agent.py loads data from a user-supplied JSON evidence file.
Boundary markers: No explicit delimiters or warnings are used for the evidence file content.
Capability inventory: The skill can execute network scans and system management commands.
Sanitization: Values from the evidence file are used solely for Boolean status counting (pass/fail) and are not interpolated into shell commands or prompt templates, effectively mitigating injection risks.
[SAFE]: The skill contains no evidence of data exfiltration, persistence mechanisms, or obfuscation. All documentation and code are consistent with professional OT security auditing practices.

implementing-nerc-cip-compliance-controls