Can I use this with other AI models besides Claude?

Yes, the Git Safety skill is compatible with both Claude Code and OpenAI Codex environments requiring git CLI access.

Why does this skill block environment file edits?

Environment files often contain sensitive secrets and local configurations; this protocol ensures only the human user manages these files to prevent accidental leaks or misconfigurations.

How does this help in team environments?

It implements coordination checks, such as verifying recent logs and requiring communication before making sweeping changes that could conflict with work done by other teammates.

Does this skill prevent Claude from making any commits?

No, it establishes a safe workflow for commits while specifically blocking destructive or unverified operations until you provide explicit written permission.

What happens if an AI agent needs to delete a file?

The skill forces the agent to stop and ask for your explicit written permission before any file is removed, even if the deletion is intended to fix a linting or build error.

Git Safety Protocol

Name: Git Safety Protocol
Author: yzlin

byyzlin

0•

开发者工具

Enforces rigorous version control safeguards and destructive operation protections to prevent data loss in collaborative development.

The Git Safety skill provides a protective framework for AI agents performing version control tasks, ensuring that high-risk commands like hard resets, force pushes, and file deletions are never executed without explicit human authorization. By implementing strict path quoting, environment file protection, and coordination-first workflows, it minimizes the risk of accidental work loss and merge conflicts in shared repositories. This skill is essential for maintaining repository integrity and security when using Claude Code or other AI-driven development tools in professional team environments.

主要功能

01Proactive coordination protocols for multi-developer or multi-agent environments

020 GitHub stars

03Strict protection for .env and sensitive environment configuration files

04Mandatory pre-commit verification and atomic commit standards

05Safety-first decision tree for handling file deletions and build error resolution

06Blocks destructive operations like hard resets and force pushes without explicit permission

使用场景

01Ensuring clean, verified, and atomic commits in production-grade team repositories

02Safeguarding sensitive environment variables from automated or unauthorized modification

03Preventing AI agents from accidentally deleting project files to resolve build or type errors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yzlin/supaviber git-safety

For use in Claude.ai and ChatGPT

Download Skill