The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill processes external infrastructure data (referred to as a 'catalog') to identify compliance violations, which then triggers automated remediation actions. This creates a high-severity vulnerability surface.
Ingestion points: GDPRComplianceChecker.check_all_requirements(catalog), HIPAAComplianceChecker.check_technical_safeguards(catalog), and SOC2ComplianceChecker.check_trust_services_criteria(catalog) in SKILL.md.
Boundary markers: No boundary markers or 'ignore embedded instructions' warnings are present to distinguish between data and potential malicious instructions within the catalog.
Capability inventory: auto_remediate_compliance_violations in SKILL.md includes capabilities to revoke_excessive_permissions, enable_encryption, apply_masking, and execute_retention_policy.
Sanitization: No evidence of sanitization or validation of the input 'catalog' data before it influences remediation decisions.
[Command Execution] (HIGH): The skill implements logic for automated administrative actions. While intended for compliance, functions like revoke_excessive_permissions and execute_retention_policy represent high-impact operations that can lead to denial of service or unauthorized access modifications if triggered maliciously through manipulated compliance reports.

compliance-automation