humanize-ai
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill is designed to process untrusted user-provided text, creating a surface for indirect instructions. 1. Ingestion points: Untrusted text provided by the user for humanization in README.md. 2. Boundary markers: Absent; there are no delimiters or instructions to ignore embedded commands within the text to be processed. 3. Capability inventory: None; no code, scripts, or tool definitions are present in the analyzed files. 4. Sanitization: Absent; no input filtering or validation logic is defined.
- [Metadata Poisoning] (LOW): There is a naming discrepancy between the _meta.json file (display name: AI Detector) and the README.md (title: Humanize-AI), which could lead to user confusion regarding the skill's purpose.
Audit Metadata