kenryu42/cc-safety-net

What it solves

It prevents AI coding agents from accidentally running destructive commands or accessing sensitive files like SSH keys and environment variables. It acts as a guardrail for your development environment, ensuring that automated tools cannot perform actions that would damage your system or leak your credentials.

How it works

It intercepts commands before they are executed by your coding agent. By parsing the intent of the command, it determines if the action is safe or dangerous. If a command is flagged as risky—such as a forced git reset or an attempt to read a private key—it blocks the execution and notifies the agent. It is not a sandbox; instead, it provides a layer of policy-based protection that works across various coding CLIs on Windows, macOS, and Linux.

Who it’s for

Developers who use AI coding agents and want to ensure these tools operate within safe boundaries without needing to manually manage complex system permissions or containerized sandboxes.

Highlights

• Blocks destructive commands like rm -rf and git push --force.
• Protects sensitive files including SSH keys, .env files, and cloud credentials.
• Provides a GUI dashboard to customize security policies and presets (Standard, Strict, or Paranoid).
• Supports team-wide security policies by committing rulebooks to your git repository.
• Offers a library API for developers to embed protection directly into their own tools.