The Agent Skills Directory

[SAFE]: The skill acts as a management layer for the Dippy permission system, utilizing standard shell commands and configuration files to audit and update local security policies. It includes robust workflows for both global and project-specific permission scopes.
[SAFE]: It incorporates critical safety defaults, such as advising against the whitelisting of interpreters (e.g., bash, sh, python3) which could be used to execute arbitrary commands and circumvent specific security rules.
[SAFE]: The skill provides a functional bash wrapper script (dippy-with-auto-fallback.sh) designed to integrate with the agent's hook system. This script implements a documented '[ASK]' convention to ensure sensitive operations always reach a human prompt even during autonomous runs.
[SAFE]: No evidence of prompt injection, data exfiltration, or obfuscation was found. The skill follows best practices for secret management, specifically instructing the use of 'deny-redirect' rules to protect credentials and environment files.

claude-permissions