humanizer
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external text while possessing file modification capabilities, making it vulnerable to indirect prompt injection. \n- Ingestion points: Text input from the user or files read via the 'Read' tool. \n- Boundary markers: Absent; the skill does not use delimiters or isolation instructions to distinguish data from instructions. \n- Capability inventory: Access to 'Write', 'Edit', 'Glob', and 'Grep' tools allows for file system modification. \n- Sanitization: Absent; no sanitization or validation of input text is provided. \n- [SAFE]: The skill's tools (Read, Write, Edit, etc.) are appropriate for its stated purpose of text editing. \n- [SAFE]: External references are limited to trusted information sources (Wikipedia).
Audit Metadata