The Agent Skills Directory

[SAFE]: The skill follows a secure design pattern by requiring explicit user confirmation before applying any proposed updates to skills or configuration files (such as CLAUDE.md). This requirement effectively mitigates the risk of unauthorized persistence of instructions derived from potentially untrusted conversation history.
[NO_CODE]: This skill consists entirely of markdown instructions for the agent and does not include any executable scripts, binaries, or package dependencies, which eliminates the technical attack surface for code-based vulnerabilities like remote code execution.

reflection