The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown-based instructions and metadata. It does not contain any executable scripts (.sh, .py, .js) or binary files, eliminating the risk of direct remote code execution or persistence.
[SAFE]: All external references in the README and SKILL files point to legitimate Wikipedia documentation and WikiProject AI Cleanup pages. There are no downloads from untrusted third-party sources.
[SAFE]: The allowed-tools list (Read, Write, Edit, Grep, Glob, AskUserQuestion) is restricted to file system operations and user interaction, which align with the skill's primary function of reviewing and editing documents.
[SAFE]: No obfuscation techniques, such as Base64 encoding, zero-width characters, or homoglyphs, were detected in any of the analyzed files.
[SAFE]: No hardcoded credentials or access patterns to sensitive system paths (e.g., .ssh, .aws) were found.
[PROMPT_INJECTION]: The skill ingests untrusted user text for humanization, which is a surface for indirect prompt injection. However, the instructions are highly specific, focusing the agent on 24 distinct linguistic patterns, which reduces the likelihood of the agent following instructions embedded within the processed text.
Ingestion points: The user-provided text for editing/humanization (referenced in SKILL.md and README.md).
Boundary markers: Absent; the skill does not explicitly instruct the agent to use delimiters or ignore instructions within the input data.
Capability inventory: File system tools (Read, Write, Edit, Grep, Glob) allow the agent to modify files based on its analysis.
Sanitization: Absent; there is no logic to escape or sanitize input content before processing.

humanizer