The Agent Skills Directory

[SAFE]: The skill provides educational templates and best practices for implementing security guards (hooks) within an agentic environment.
[SAFE]: It employs robust parsing of shell inputs using Python's shlex module to prevent command injection and common filter bypasses that rely on shell quoting or tokenization errors.
[SAFE]: The documentation explicitly identifies and provides mitigations for high-risk behaviors, such as non-terminating loops in blocking hooks and silent failures in state-deriving guards.
[SAFE]: It promotes secure interaction patterns, specifically encouraging human-confirmation gates for irreversible actions (Tier-0) that cannot be controlled via software-only mechanisms.
[SAFE]: No evidence of malicious intent, unauthorized data access, or network exfiltration was found in the provided templates and documentation.

claude-code-hooks