Agency
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill package is composed exclusively of Markdown files for templates and guidance. It contains no Python scripts, JavaScript files, shell scripts, or binary files.
- [SAFE]: A comprehensive audit of all files against the ten threat categories revealed no malicious patterns, obfuscation, or safety violations.
- [PROMPT_INJECTION]: No adversarial instructions or jailbreak attempts were found. The skill reinforces safety by requiring human approval for client-facing communications and financial proposals.
- [DATA_EXFILTRATION]: While the skill manages sensitive agency information, it does so within a local workspace (typically ~/agency/). There are no tools, network operations, or scripts present that could facilitate unauthorized data transmission.
Audit Metadata