jj-workflow

Jujutsu (jj) version control for AI-assisted development. Use when:- Repo has `.jj/` directory (check with `jj root`)- Need undo/recovery after breaking changes- Curating messy history before PR

⭐5stars🔀1forks📁View on GitHub🕐Updated Jan 11, 2026

When & Why to Use This Skill

This Claude skill integrates Jujutsu (jj), a next-generation version control system, into AI-assisted development workflows. It simplifies versioning by eliminating the staging area, providing a robust operation log for instant recovery from breaking changes, and allowing developers to curate clean, professional commit histories before pushing to Git repositories. It is specifically designed to provide a safety net for rapid, AI-driven iterations.

Use Cases

  • Instant Recovery: Use the operation log to immediately revert to a functional state if an AI-generated code change introduces breaking errors or unexpected behavior.
  • History Curation: Seamlessly combine and organize messy incremental checkpoints into logical, readable commits using squash and describe commands before submitting a Pull Request.
  • Risk-Free Experimentation: Quickly create new change branches to test experimental features or refactors, with the ability to easily abandon them if they fail without impacting the main codebase.
  • Git-Compatible Workflow: Manage complex local versioning logic using jj's mutable change model while maintaining full compatibility with standard Git-based team environments.
namejj-workflow
description|

jj Workflow

Mental Model

No staging area. Your working directory is always a commit. Every save is tracked.

  • @ = your current change (the working copy)
  • @- = parent of current change
  • Changes are mutable until pushed

When to Use What

Situation Do This
Starting new work jj new -m "what I'm trying"
Work is done, move on jj new (current becomes parent)
Annotate what you did jj describe -m "feat: auth"
Broke something jj op log → jj op restore <id>
Undo one file jj restore --from @- <path>
Combine messy commits jj squash
Split bloated commit jj split
Try something risky jj new -m "experiment", then jj abandon @ if it fails

AI Coding Pattern

During implementation, don't think about commits. Just work. jj tracks everything.

# Before risky change
jj describe -m "checkpoint: auth works"
jj new -m "adding OAuth"

# If it breaks
jj op log              # Find good state
jj op restore <id>     # Go back

# When done, curate
jj log -r 'ancestors(@, 10)'
jj squash              # Combine checkpoints
jj describe -m "feat: OAuth support"

Push to GitHub

jj bookmark create feature-x   # Name for pushing
jj git push --allow-new        # Push to remote

Teammates see clean git. They don't know you used jj.

Recovery

The oplog records every operation. Nothing is lost.

jj op log                      # See all operations
jj undo                        # Undo last operation
jj op restore <id>             # Jump to any past state

Bail Out

rm -rf .jj    # Delete jj, keep git unchanged