The Agent Skills Directory

[SAFE]: The skill's instructions establish a secure framework for agent autonomy by distinguishing between low-risk reversible actions (e.g., writing code or notes) and high-risk irreversible actions (e.g., deleting production data or sending messages). It explicitly mandates human authorization for the latter, which aligns with safety best practices.
[SAFE]: External references in the documentation point to the official GitHub repository for Cursor plugins, a well-known resource in the AI development community. These links are for attribution and maintenance purposes and do not involve runtime remote code execution or suspicious dependency management.

pstack-principle-never-block-on-the-human