parallel-worktree - Parallel Worktree Orchestration for Stacked Implementation
Reads a plan file and uses worktrunk (wt) to split worktrees, orchestrating parallel and stacked implementation with tmux-based Claude sessions and optional Remote Control.
Tags
Updated: 2026-06-30Capabilities
Typical Inputs
Typical Outputs
What this skill does
- read plan file
- generate JSON task spec
- detect task dependencies
- run preflight checks
- calculate parallel scheduling waves
- generate wt and tmux commands
- create git worktrees via wt
- launch parallel tmux sessions
- build stacked branches
- push feature branches
- create draft pull requests
- monitor worktree status
- remove merged worktrees
Inputs
- Plan file path
- wt (worktrunk) CLI
- gh CLI
- tmux
- uv Python package manager
- Git repository
- Remote Control option flag
Outputs
- Git worktrees
- Parallel tmux sessions with Claude agents
- Feature branches
- Draft pull requests
- Schedule and command execution sequences
Requirements
- wt (worktrunk) must be installed
- gh CLI must be installed
- tmux must be available
- uv must be installed for Python environment
- Git repository required
- Explicit user invocation only (disable-model-invocation: true)
- Plan file must be pre-created before invocation
- commit-flow and pr-create skills available