humanize
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructions use strong, imperative language to override standard agent behavior. It claims to be a "MANDATORY skill" and the "ONLY tool allowed" for specific writing tasks, which attempts to constrain the agent's default decision-making process.
- [PROMPT_INJECTION]: The skill provides explicit instructions to bypass or erase "model-specific fingerprints" from Claude, GPT, and Gemini, which is a form of instruction meant to alter the fundamental output characteristics of the underlying models.
- [INDIRECT_PROMPT_INJECTION]: This skill is designed to process untrusted data (user-supplied text) to rewrite it.
- Ingestion points: User-provided text to be "humanized" as described in SKILL.md.
- Boundary markers: None identified in the instructions to prevent the agent from obeying instructions embedded within the text being rewritten.
- Capability inventory: No dangerous capabilities (subprocess execution, file system writes, or network operations) are present in the skill.
- Sanitization: No sanitization or filtering of the input text is described.
Audit Metadata