git-guardrails-claude-code

The git-guardrails-claude-code skill is designed to enhance safety when using Claude Code by intercepting and blocking potentially destructive Git commands before they execute. By setting up PreToolUse hooks, this skill prevents operations such as forced pushes, hard resets, branch deletions, and file cleanups that could result in loss of work or repository corruption. It provides an additional layer of protection, ensuring that users do not accidentally execute commands that could compromise project integrity, especially in collaborative or production environments.

The main capabilities of this skill include blocking specific dangerous Git commands such as `git push` (including `--force`), `git reset --hard`, `git clean -f` or `git clean -fd`, `git branch -D`, and `git checkout .` or `git restore .`. It provides clear feedback to Claude Code whenever a blocked command is attempted, notifying it that it does not have permission to execute the command. The skill also allows for flexible installation, either on a per-project basis or globally across all projects. Users can customize the blocked command list, merging their preferences into the existing hook script without overwriting other settings, and verify functionality through a simple test command.

This skill is particularly useful for developers, DevOps engineers, and teams working with Git in Claude Code who want to enforce safe practices and prevent accidental destructive operations. It is ideal for scenarios where repository safety is critical, such as shared development environments, open-source contributions, or production deployments. By implementing these guardrails, users can confidently interact with Git, reducing the risk of accidental data loss while maintaining workflow efficiency.