A safety-first workflow for AI-assisted coding A safety-first workflow for AI-assisted coding TL;DR: Commit your code before asking an AI Assistant to change it. TL;DR: Commit your code before asking an AI Assistant to change it. Common Mistake ❌ Developers ask the AI assistant to "refactor this function" or "add error handling" while they have uncommitted changes from their previous work session. When the AI makes its changes, the git diff shows everything mixed together—their manual edits plus the AI's modifications. If something breaks, they can't easily separate what they did from what the AI did and make a safe revert. Problems Addressed 😔 You mix your previous code changes with AI-generated code.
You lose track of what you changed.
You struggle to revert broken suggestions. You mix your previous code changes with AI-generated code. You lose track of what you changed. You struggle to revert broken suggestions. How to Do It 🛠️ Finish your manual task.
Run your tests to ensure everything passes.
Commit your work with a clear message like "feat: manual implementation of X".
You don't need to push your changes.
Send your prompt to the AI assistant.
Review the changes using your IDE's diff tool.
Accept or revert: Keep the changes if they look good, or run git reset --hard HEAD to instantly revert.
Run the tests again to verify AI changes didn't break anything.
Commit AI changes separately with a message like refactor: AI-assisted improvement of X. Finish your manual task. Run your tests to ensure everything passes. Commit your work with a clear message like "feat: manual implementation of X". You don't need to push your changes. Send your prompt to the AI assistant. Review the changes using your IDE's diff tool. Accept or revert: Keep the changes if they look good, or run git reset --hard HEAD to instantly revert. Accept or revert git reset --hard HEAD Run the tests again to verify AI changes didn't break anything. Commit AI changes separately with a message like refactor: AI-assisted improvement of X. refactor: AI-assisted improvement of X Benefits 🎯 Clear Diffing: You see the AI's "suggestions" in isolation. Easy Revert: You can undo a bad AI hallucination instantly. Context Control: You ensure the AI is working on your latest, stable logic. Tests are always green: You are not breaking existing functionality. Context 🧠 When you ask an AI to change your code, it might produce unexpected results. It might delete a crucial logic gate or change a variable name across several files. If you have uncommitted changes, you can't easily see what the AI did versus what you did manually. When you commit first, you create a safety net. You can use git diff to see exactly what the AI modified. git diff If the AI breaks your logic, you can revert to your clean state with one command. You work in very small increments. very small increments Some assistants are not very good at undoing their changes. undoing Prompt Reference 📝 git status              # Check for uncommitted changes

git add .               # Stage all changes

git commit -m "msg"     # Commit with message

git diff                # See AI's changes

git reset --hard HEAD   # Revert AI changes

git log --oneline       # View commit history git status              # Check for uncommitted changes

git add .               # Stage all changes

git commit -m "msg"     # Commit with message

git diff                # See AI's changes

git reset --hard HEAD   # Revert AI changes

git log --oneline       # View commit history Considerations ⚠️ This is only necessary if you work in write mode and your assistant is allowed to change the code. write mode Type 📝 [X] Semi-Automatic You can enforce the rules of your assistant to check the repository status before making changes. Limitations ⚠️ If your code is not under a source control system, you need to make this manually. Tags 🏷️ Complexity Complexity Level 🔋 [X] Beginner Related Tips 🔗 Use TCR
Practice Vibe Test Driven Development
Break Large Refactorings into smaller prompts
Use Git Bisect for AI Changes: Using git bisect to identify which AI-assisted commit introduced a defect
Reverting Hallucinations Use TCR TCR Practice Vibe Test Driven Development Break Large Refactorings into smaller prompts Use Git Bisect for AI Changes: Using git bisect to identify which AI-assisted commit introduced a defect Git Bisect git bisect Reverting Hallucinations Conclusion 🏁 Treating AI as a pair programmer requires the same safety practices you'd use with a human collaborator: version control, code review, and testing. When you commit before making a prompt, you create clear checkpoints that make AI-assisted development safer and more productive. This simple habit transforms AI from a risky black box into a powerful tool you can experiment with confidently, knowing you can always return to a working state. Commit early, commit often, and don't let AI touch uncommitted code. More Information ℹ️ https://hackernoon.com/git-explained-in-5-levels-of-difficulty?embedable=true https://hackernoon.com/git-explained-in-5-levels-of-difficulty?embedable=true https://www.infoq.com/articles/test-commit-revert/?embedable=true https://www.infoq.com/articles/test-commit-revert/?embedable=true https://medium.com/@kentbeck_7670/test-commit-revert-870bbd756864?embedable=true https://medium.com/@kentbeck_7670/test-commit-revert-870bbd756864?embedable=true Tools 🧰 GIT is an industry standard, but you can apply this technique to any other version control software. GIT This article is part of the AI Coding Tip Series. AI Coding Tip

Keep

Code Smell 12 - Null is Schizophrenic and Does Not Exist in The Real-world

Buy My Book

Share your ideas with me!

AI Coding Tip 001 - Commit Your Code Before Asking For Help From an AI Assistant

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

10 Predictions About the Future of Finance

AI-Assisted Coding with Tabnine

DOE v. Github: Code Stored on GitHub Ultimately Belong to Their Authors

DOE v. Github: Defendants' Conduct Raises Questions Related to DMCA Violations, Privacy, Data Breach

DOE v. Github: Defendants Never Sought Permission to Use Others' Code to Train Codex or Copilot

DOE v. Github: Lawsuit Falls Under California's Jurisdiction Because of GitHub & OpenAI's HQ

10 Predictions About the Future of Finance

AI-Assisted Coding with Tabnine

DOE v. Github: Code Stored on GitHub Ultimately Belong to Their Authors

DOE v. Github: Defendants' Conduct Raises Questions Related to DMCA Violations, Privacy, Data Breach

DOE v. Github: Defendants Never Sought Permission to Use Others' Code to Train Codex or Copilot

DOE v. Github: Lawsuit Falls Under California's Jurisdiction Because of GitHub & OpenAI's HQ

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps