The Agent Skills Directory

[SAFE]: The skill serves as a shared foundation for lark-cli interactions, emphasizing strict safety rules. It explicitly instructs the agent to avoid outputting secrets, mandates explicit user confirmation for high-risk write or delete actions (handling exit code 10), restricts file path access to relative paths to avoid directory traversal, and enforces the use of structured argument arrays (argv) instead of shell command concatenation to prevent shell injection vulnerabilities. No threat patterns or malicious indicators were found across any of the files.

lark-shared