humanize
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill implements strict content integrity rules ('The Cardinal Rule') that forbid the agent from fabricating anecdotes, statistics, or citations, ensuring that only stylistic transformations are performed.
- [COMMAND_EXECUTION]: The TESTING.md file contains common Unix shell commands (wc, grep, awk, tr, sort) which are intended for objective measurement of text metrics like word count and sentence length variance during the testing process.
- [PROMPT_INJECTION]: Analysis of the skill's instructions shows no attempts at prompt injection, data exfiltration, or obfuscation; all 'CRITICAL' directives are task-specific quality constraints used to enforce stylistic consistency.
Audit Metadata