The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional text and metadata. It does not include executable scripts, external dependencies, or requests for elevated system privileges.
[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external content (text provided by users or read from files), which is a common surface for indirect prompt injection. This is an inherent risk of the skill's primary function.
Ingestion points: Text strings provided in the prompt and files accessed through the read_file tool (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the text being processed.
Capability inventory: The skill is intended to use read_file, patch, and write_file to perform text transformations (SKILL.md).
Sanitization: No sanitization or validation of the input text is prescribed beyond identifying linguistic patterns.

humanizer