> ## Documentation Index
> Fetch the complete documentation index at: https://ade-ac1c6011-dependabot-github-actions-actions-cache-6.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.
# Multi-agent setup
> Run several agents safely by giving each task its own lane, watching for collisions, and choosing the grid when you want to compare attempts.
The safest multi-agent setup is simple: one lane per task, one clear goal per agent, and a review pass before merge. Isolation is what keeps parallel agents from writing over each other — give each its own worktree and the collisions disappear.
## One lane per task
Every lane is its own git worktree — its own branch, files, and sessions. That isolation is the whole game: two agents in two lanes can't touch each other's working tree, so they run in parallel without stashing, rebasing, or fighting over the same files.
Turn a large goal into small, independent, lane-sized tasks.
Name lanes after the outcome, not the agent.
Keep each prompt scoped to that lane's goal.
Use Work for live sessions and Graph for branch relationships and merge risk.
Inspect each diff, run tests, and open PRs separately.
## Good task splits
A clean split has slices that touch different files and can be reviewed on their own.
| Goal | Better split |
| ------------------------- | ---------------------------------------------------------------------- |
| "Improve onboarding" | Copy fixes, provider validation, first-run UI, tests. |
| "Fix checkout bugs" | Repro the failing case, patch backend, patch UI, add regression tests. |
| "Ship mobile sync polish" | Desktop status, iOS display, reconnect behavior, docs. |
## Watch for collisions
Parallel agents stay safe as long as their work doesn't overlap. Most collisions come from scope creep, not from the tool.
* Don't send two agents into the same lane for unrelated work.
* Avoid broad prompts that invite repo-wide edits — keep each agent narrow.
* Use [child lanes](/lanes/stacks) when one task genuinely depends on another, rather than racing them.
* Merge parent work before rebasing children when possible.
* Use History to see exactly what each session changed.
## Grid vs. separate lanes
Both run agents in parallel, but they answer different questions.
Same task, several models — race a narrow, well-specified change and keep the best diff. Each tile is a full chat.
Different tasks — give each its own worktree so broad, multi-step work never competes for the same files.
Use the grid for narrow, well-specified work where comparing a few attempts is cheap. For broad, multi-step features, give each agent its own lane instead.
Run several agents at once and compare side by side.
Isolate each task in its own worktree.