Show HN: A 60-line PreToolUse hook that stops Claude Code from editing your .env (github.com)

A new tool named "guard.py" has been introduced to prevent Claude Code from unintentionally editing sensitive configuration files, specifically the `.env` file. This 60-line script operates before any tool calls, effectively blocking any attempts to modify files that match pre-defined patterns listed in a protected file. By implementing this pre-tool hook, developers can enforce stricter safeguards, ensuring that critical files are protected from automatic edits by the AI model, which often loses adherence to its guidelines during longer interactions. This development is significant for the AI/ML community as it enhances security measures during AI-assisted coding, mitigating potential risks of inadvertent data exposure or corruption. The guard not only prevents undesired edits but also provides feedback on why a particular operation was blocked, enabling a smoother workflow. The tool is straightforward to install, requires no additional dependencies, and is designed to operate securely alongside existing workflows, thus promoting safer use of AI coding assistants. This initiative is part of a broader effort towards developing robust safety measures in AI applications, establishing a crucial reference point for responsible AI development practices.