The Agent Skills Directory

[SAFE]: The skill provides legitimate instructions for editing and humanizing text. It contains strong constraints against hallucination ("Never invent facts") and specifies clear boundaries for its operation.
[SAFE]: The skill references an external repository on GitHub (github.com/cursor/plugins) as a source for its pattern catalog. This reference is to a well-known service and does not involve downloading or executing external scripts.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted text from user input or file paths. While this creates a surface for indirect prompt injection, the skill mitigates this by instructing the agent to focus only on prose transformation and to leave structured data, code blocks, and frontmatter untouched.

humankit