The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is explicitly designed to act as a security layer that gates high-impact operations such as file system destruction, infrastructure changes, and credential modification, requiring explicit user authorization before execution.
[SAFE]: The audit.py utility utilizes yaml.safe_load() for reading state files, which is a recommended security practice to prevent arbitrary code execution via YAML deserialization attacks.
[SAFE]: The skill maintains all operational state and logs locally within the user's home directory (~/.openclaw), with no evidence of unauthorized network access or external data transmission.

dangerous-action-guard