Worktrees: Isolated Agent Workspaces

Git worktrees let you run multiple Claude Code sessions on the same repo without stepping on each other's diffs. They're the underrated unlock for parallel agent work.

9 min · Reviewed 2026

What worktrees solve

Two Claude Code sessions in the same checkout will fight. They both see the same files, both edit them, and your diff becomes a salad. Worktrees give each session its own working directory backed by the same git repo, so each can have its own branch and its own state without conflict.

The basic move

# From your main checkout git worktree add ../my-project-feat-auth feat-auth git worktree add ../my-project-refactor-api refactor-api # Run a Claude session in each (cd ../my-project-feat-auth && claude) (cd ../my-project-refactor-api && claude) # When done git worktree remove ../my-project-feat-authEach worktree is a separate folder backed by the same repo. Clean up when the work merges.

Where worktrees shine

Two features in flight, two agents progressing each one
Spawning a 'spike' agent on an experimental branch while you keep working on main
Reviewing a teammate's PR with an agent without leaving your active branch
Hot-fix on top of a long-running refactor, neither blocking the other
Trying three different approaches to the same problem, one per worktree

Apply: spin up one parallel session

Pick a feature branch you're partway through
Create a worktree on a different branch
Run a separate Claude Code session in it
Make sure both sessions are working on truly independent files
Merge or remove the worktree when the work lands

The big idea: one agent, one worktree. The git worktree command turns chaotic parallelism into clean parallelism.

End-of-lesson check

8 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-claude-code-worktrees-creators

What is the main idea of "Worktrees: Isolated Agent Workspaces"?
1. Git worktrees let you run multiple Claude Code sessions on the same repo without stepping on each other's diffs.
2. Use AI as the final authority for the whole decision
3. Avoid checking the answer once it sounds polished
4. Focus only on speed instead of judgment
Which concept is most central to "Worktrees: Isolated Agent Workspaces"?
1. branch isolation
2. worktree
3. parallel sessions
4. merge conflict
Which use of AI fits this topic best?
1. Let the AI decide what matters without your review
2. Use the answer before checking whether it fits the situation
3. Two features in flight, two agents progressing each one
4. Treat the AI output as automatically correct
What should a careful learner remember about "One agent, one worktree"?
1. Use AI to draft or organize ideas about worktree, then verify before acting.
2. Skip the context so the tool can guess faster
3. Treat the output as private even after sharing it online
4. Use the answer without checking the source
You want to use AI after this lesson. What is the safest next step?
1. Act immediately because the AI answer is written clearly
2. Use AI for drafting and comparison, but verify before publishing or relying on it.
3. Hide uncertainty so the final answer looks cleaner
4. Use private or sensitive details before checking permission
How should AI output about worktree be treated?
1. As proof that no other source is needed
2. As a replacement for context, consent, or expert review
3. As a draft or helper output that still needs human judgment and verification
4. As something that becomes correct when it sounds confident
Name one way to verify an AI answer about worktree.
Which action would help you apply "Worktrees: Isolated Agent Workspaces" responsibly?
1. Use the tool to avoid thinking through the tradeoff
2. Keep going even if the output conflicts with a trusted source
3. Treat the AI output as automatically correct
4. Spawning a 'spike' agent on an experimental branch while you keep working on main

← Back to interactive lesson

Tendril · Creators · Tools Literacy

Worktrees: Isolated Agent Workspaces

Git worktrees let you run multiple Claude Code sessions on the same repo without stepping on each other's diffs. They're the underrated unlock for parallel agent work.

9 min · Reviewed 2026

What worktrees solve

The basic move

# From your main checkout git worktree add ../my-project-feat-auth feat-auth git worktree add ../my-project-refactor-api refactor-api # Run a Claude session in each (cd ../my-project-feat-auth && claude) (cd ../my-project-refactor-api && claude) # When done git worktree remove ../my-project-feat-authEach worktree is a separate folder backed by the same repo. Clean up when the work merges.

Where worktrees shine

Two features in flight, two agents progressing each one
Spawning a 'spike' agent on an experimental branch while you keep working on main
Reviewing a teammate's PR with an agent without leaving your active branch
Hot-fix on top of a long-running refactor, neither blocking the other
Trying three different approaches to the same problem, one per worktree

Apply: spin up one parallel session

Pick a feature branch you're partway through
Create a worktree on a different branch
Run a separate Claude Code session in it
Make sure both sessions are working on truly independent files
Merge or remove the worktree when the work lands

The big idea: one agent, one worktree. The git worktree command turns chaotic parallelism into clean parallelism.

End-of-lesson check

8 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-claude-code-worktrees-creators

What is the main idea of "Worktrees: Isolated Agent Workspaces"?
1. Git worktrees let you run multiple Claude Code sessions on the same repo without stepping on each other's diffs.
2. Use AI as the final authority for the whole decision
3. Avoid checking the answer once it sounds polished
4. Focus only on speed instead of judgment
Which concept is most central to "Worktrees: Isolated Agent Workspaces"?
1. branch isolation
2. worktree
3. parallel sessions
4. merge conflict
Which use of AI fits this topic best?
1. Let the AI decide what matters without your review
2. Use the answer before checking whether it fits the situation
3. Two features in flight, two agents progressing each one
4. Treat the AI output as automatically correct
What should a careful learner remember about "One agent, one worktree"?
1. Use AI to draft or organize ideas about worktree, then verify before acting.
2. Skip the context so the tool can guess faster
3. Treat the output as private even after sharing it online
4. Use the answer without checking the source
You want to use AI after this lesson. What is the safest next step?
1. Act immediately because the AI answer is written clearly
2. Use AI for drafting and comparison, but verify before publishing or relying on it.
3. Hide uncertainty so the final answer looks cleaner
4. Use private or sensitive details before checking permission
How should AI output about worktree be treated?
1. As proof that no other source is needed
2. As a replacement for context, consent, or expert review
3. As a draft or helper output that still needs human judgment and verification
4. As something that becomes correct when it sounds confident
Name one way to verify an AI answer about worktree.
Which action would help you apply "Worktrees: Isolated Agent Workspaces" responsibly?
1. Use the tool to avoid thinking through the tradeoff
2. Keep going even if the output conflicts with a trusted source
3. Treat the AI output as automatically correct
4. Spawning a 'spike' agent on an experimental branch while you keep working on main

← Back to interactive lesson