humanizer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions are focused entirely on stylistic text editing and linguistic improvements. No patterns were found for prompt injection, unauthorized data access, or persistence mechanisms. The installation instructions provided in the README are standard for CLI tools and point to a reputable host (GitHub).
- [NO_CODE]: The skill consists exclusively of Markdown instructions and YAML configuration. There are no Python scripts, Node.js files, shell scripts, or binary executables included in the package, eliminating the risk of direct malicious code execution.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process text provided by the user to remove AI artifacts. This creates an input surface where malicious instructions could potentially be embedded in the text to be humanized.
- Ingestion points: User-provided text strings passed to the agent via the /humanizer command or direct prompts.
- Boundary markers: The instructions do not explicitly define delimiters to separate user text from instructions, though the specific focus on pattern matching reduces the risk of obedience to embedded commands.
- Capability inventory: The skill has access to file system tools (Read, Write, Edit, Grep, Glob) and the ability to ask the user questions.
- Sanitization: No specific input sanitization or "ignore embedded instructions" directives are present in the SKILL.md instructions.
Audit Metadata