Use local for direct hands-on debugging, worktrees for isolated parallel changes, and cloud for remote or background execution.
The decision behind the technique
The environment changes which files, services and uncommitted changes the task can see. A wrong choice creates missing dependencies, branch collisions or accidental coupling to your main checkout.
Use local for direct hands-on debugging, worktrees for isolated parallel changes, and cloud for remote or background execution.
A practical step-by-step workflow
Use this sequence as a default, then tighten or expand it only when the evidence changes.
- Identify whether uncommitted local state is required
- Choose a worktree when parallel edits need isolation
- Choose cloud only when the task can reproduce its environment remotely
- Verify branch, directory and dependencies before editing
Failure boundary and finish line
A worktree does not automatically copy ignored configuration, local databases or running services.
Finish when the next action, owner, evidence and stop condition are all explicit.
Frequently asked questions
What should I do first?+
Identify whether uncommitted local state is required
What is the main failure mode?+
A worktree does not automatically copy ignored configuration, local databases or running services.
OFFICIAL SOURCES
Verify against the source
OpenAI worktree guidanceOpenAI local environmentsThis guide provides general information, not an OpenAI allowance promise or support decision. Your account and official pages govern actual usage, benefits and recovery times.