The Agent Skills Directory

[SAFE]: The skill serves as a set of instructions for refining text. It does not perform any unauthorized network activities, access sensitive credentials, or attempt to exfiltrate data.- [PROMPT_INJECTION]: The skill processes user-supplied text to perform editorial tasks. Ingestion points: user text input in Claude Code. Boundary markers: absent. Capability inventory: Read, Write, Edit, Grep, Glob tools. Sanitization: absent. While this represents a surface for indirect prompt injection, it is standard for text editing skills and no malicious exploitation patterns are present.- [COMMAND_EXECUTION]: The tools requested in the YAML frontmatter (Read, Write, Edit, Grep, Glob) are standard for file manipulation in Claude Code. The skill uses these tools only to modify user-provided text as requested.- [EXTERNAL_DOWNLOADS]: The documentation mentions downloading the skill from a GitHub repository for installation. This is a standard and safe deployment method for skills in this environment.

humanizer