TaskPods: Isolated Parallel AI Agent Workflows

Описание и применение

TaskPods позволяет запускать несколько AI‑агентов в изолированных git‑рабочих деревьях, автоматически управляя ветками, PR‑ами и очисткой.

Multiple AI coding agents editing the same repository caused conflicting changes and manual git worktree management.

TaskPods enables isolated, parallel AI agent workflows without manual worktree juggling, streamlining code changes, reviews, and PR creation.

Install TaskPods via pip
Start a pod with taskpods start <pod-name> --agent "<agent-command>"
Agent runs in its dedicated worktree and branch
Review the agent’s diff with git -C .taskpods/<pod-name> diff main
If acceptable, finalize with taskpods done <pod-name> -m "<commit-message>" --remove
If not, abort with taskpods abort <pod-name>

Two agents on the same repo means they edit the same files, stage each other's changes, and occasionally commit on top of one another.

That creates .taskpods/fix-auth/ as a worktree on branch pods/fix-auth and runs the agent there.

taskpods done fix-auth -m "fix: token refresh" --remove