humanize
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prose text to identify and remove AI-writing patterns.
- Ingestion points: User-provided draft text in the agent's context for humanization.
- Boundary markers: None explicitly defined in the instructions; the skill relies on the agent's default prompt boundaries.
- Capability inventory: The skill is limited to text transformation and analysis; it does not utilize any shell commands, file system operations, or network tools.
- Sanitization: No input sanitization or filtering is applied to the ingested prose.
- [EXTERNAL_DOWNLOADS]: The skill references various research papers, datasets, and repositories to establish baselines for AI-writing metrics.
- Evidence: The skill cites sources from ArXiv, Wikipedia, and public repositories (e.g.,
osolmaz/ai-smell) for documentation and attribution purposes. These references are to well-known academic and open-source platforms.
Audit Metadata