skill-system-foundry
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides local Python entry points (validate_skill.py, scaffold.py, bundle.py, audit_skill_system.py) that handle filesystem operations such as reading, writing, and zipping skill components. These scripts use only the Python standard library and are scoped strictly to the lifecycle management of skills within a user-defined directory.
- [PROMPT_INJECTION]: No evidence of bypass markers, behavior override instructions, or jailbreak role-play was detected in the skill's instructions or templates. The meta-skill follows best practices for progressive disclosure.
- [DATA_EXFILTRATION]: Analysis of the Python library modules (lib/discovery.py, lib/reachability.py) confirms that file access is confined to identifying and validating internal references. No unauthorized network operations or harvesting of sensitive system files (~/.ssh, ~/.aws) were found.
- [REMOTE_CODE_EXECUTION]: The skill logic is entirely self-contained. It does not perform remote downloads of scripts or execute unverifiable packages at runtime. External links point to trusted documentation or the author's official GitHub repository.
Audit Metadata