The Agent Skills Directory

[SAFE]: The skill implements a robust safety model for infrastructure management. All operations that modify network state are permission-gated through environment variables and require a two-step 'preview-then-confirm' process to ensure user oversight and prevent accidental misconfiguration.
[SAFE]: Sensitive information is handled securely. The skill defines clear rules for redacting secrets (passphrases, SSH credentials, private keys) from tool outputs, ensuring they are not exposed in the agent's context or logged as plain text.
[SAFE]: The tool definitions and usage guidelines provide clear warnings about the impact of high-risk operations, such as rebooting devices or modifying radio settings, which helps ensure stable network operations.

unifi-network