The Agent Skills Directory

[SAFE]: The skill is designed for text rewriting and style alignment. It provides a structured methodology for identifying and removing typical AI writing patterns through a multi-step diagnostic and rewriting process without executing any dangerous commands or scripts.
[DATA_EXPOSURE]: The skill reads configuration from ./writing-workspace/styles/my_style.json. This is a standard and safe practice for loading user-defined style profiles within a dedicated workspace and does not involve accessing sensitive system credentials, private SSH keys, or environment variables.
[PROMPT_INJECTION]: The skill processes untrusted user input (articles to be rewritten), which is a common surface for indirect prompt injection. However, since the skill does not use any high-privilege tools such as network access, shell execution, or arbitrary file writing, any attempted injection would be confined to the text output of the current session.

rewrite