The Agent Skills Directory

[SAFE]: The skill is entirely composed of Markdown instructions and prompt templates. No scripts (Python, Node.js, Shell) or binary executables are included, eliminating the risk of unauthorized code execution.
[SAFE]: File access is restricted to reading local project files (ABOUT-ME.md and MY-ICP.md) for the purpose of adopting the user's personal writing style and understanding their target audience. This is a standard and safe operation for a writing assistant.
[SAFE]: No obfuscation, hidden URLs, or credential-harvesting patterns were identified. All instructions are transparent and aligned with the skill's stated purpose of prose generation.
[SAFE]: The skill does not perform any network requests or external data transfers. All interactions remain within the agent's standard conversational context.

prose-writing