The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of shell commands generated by AI models. It explicitly promotes safety by encouraging the use of sandboxed environments (e.g., --sandbox read-only) and clearly marking flags that bypass safety protocols.
[SAFE]: All external models and providers referenced (DeepSeek, SiliconFlow, Ollama, etc.) are well-known AI services. The skill interacts with these services through the author's own x-cmd CLI utility, with no evidence of unauthorized data exfiltration or suspicious network activity.
[SAFE]: The metadata and functional descriptions are consistent with the skill's code generation and terminal agent capabilities. No obfuscation, persistence mechanisms, or privilege escalation patterns were identified.

codex