The Agent Skills Directory

[SAFE]: The skill provides comprehensive instructions for identifying and rewriting AI-generated text patterns. Its logic is based on public guidelines from a well-known service (Wikipedia's WikiProject AI Cleanup) and focuses on stylistic improvements.
[SAFE]: All referenced tools (Read, Write, Edit, Grep, Glob, AskUserQuestion) are appropriate for the skill's function as a writing editor. No unauthorized network access, privilege escalation, or dangerous command executions were detected.
[SAFE]: The skill processes external text and files for voice calibration and humanization. While this constitutes an indirect prompt injection surface, it is inherent to the skill's utility and no malicious exploitation patterns were found.
Ingestion points: Content provided by the user for processing and reference file paths for style matching defined in SKILL.md.
Boundary markers: The skill relies on natural language instructions for context and does not utilize specific technical delimiters for input data.
Capability inventory: File system access via Read, Write, and Edit tools in SKILL.md.
Sanitization: No explicit sanitization or filtering of input text is performed before processing.
[SAFE]: No obfuscation techniques such as base64, hidden characters, or homoglyphs were found. No hardcoded credentials or sensitive file path access (outside of user-directed files) were identified.

humanizer