general-skill-refiner
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [SAFE] (SAFE): Extensive review of the content shows it is a documentation file providing guidelines for skill quality. No prompt injection, data exfiltration, or obfuscation attempts were found.
- [COMMAND_EXECUTION] (LOW): The document provides utility command examples using
grepandwc. These are used to analyze local file lengths and find specific keywords (like time estimates) within the skill directory. These commands do not involve network access, privilege escalation, or modification of system files. - [DATA_EXPOSURE] (INFO): The bash examples reference relative paths (e.g.,
../critical-app-brief/) to compare metrics against benchmarks. This is a standard practice for development environments and does not target sensitive user data or system credentials.
Audit Metadata