humanizer
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown-based instructions and examples, with no executable scripts, binaries, or shell commands provided.
- [SAFE]: The skill references a well-known, trusted source (Wikipedia's 'Signs of AI writing' guide) for its core functionality.
- [SAFE]: All allowed tools (
Read,Write,Edit,Grep,Glob,AskUserQuestion) are appropriate for the skill's primary purpose as a writing assistant. - [PROMPT_INJECTION]: The skill processes untrusted input text, creating a surface for indirect prompt injection. Ingestion points: input text for humanization provided by the user or read from files. Boundary markers: none specified to isolate the input from the agent's instructions. Capability inventory: standard file system tools and user interaction. Sanitization: none described. The risk is considered safe because the skill performs only text editing and lacks network or administrative capabilities.
Audit Metadata