The Agent Skills Directory

[SAFE]: The skill's instructions and configuration are consistent with its stated purpose as a security scanner. No obfuscation, data exfiltration, or malicious commands were detected.
[PROMPT_INJECTION]: The skill processes untrusted data from other skills and agents, which creates a potential surface for indirect prompt injection. A malicious file being scanned could theoretically attempt to influence the auditor's output.
Ingestion points: Scans files located in app/skills/ and app/agents/.
Boundary markers: No explicit delimiters or boundary markers are defined in the instructions to isolate the content being scanned.
Capability inventory: Utilizes Read, Grep, Glob, and Bash tools, and executes a local Python script scripts/audit_skills.py to perform the audit.
Sanitization: The skill does not specify any sanitization procedures for the external content it processes.

skill-audit