The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a Bash script (dippy-with-auto-fallback.sh) and instructions to configure it as a PreToolUse hook in the agent's settings.json. This script acts as a proxy for the Dippy permission manager to filter prompts during autonomous operations.
[SAFE]: The skill manages the agent's internal security configuration files (~/.dippy/config, .dippy, settings.json, and .claude/settings.local.json). This is the core functionality and is used to define the boundaries of the agent's capabilities.
[SAFE]: The skill documentation and the provided config.example include security best practices, such as deny-redirect rules to prevent the agent from writing to sensitive files like .env, SSH keys (id_rsa), and PEM files.
[SAFE]: The skill implements a human-in-the-loop requirement by recommending ask or deny rules for destructive operations and uses a specific token [ASK] to ensure critical prompts reach the user even in autonomous modes.

agent-permissions