- Home
- Skills
- Agents & Orchestration
- multi-agent-worktree-orchestrator
Works with the AI Agents you already use
Multi Agent Worktree Orchestrator
Orchestrate parallel coding agents without git index corruption, build-lock contention, or token-wasting ping-pong using isolated worktrees and scoped briefs.
$19
multi-agent-worktree-orchestrator
Example session with this skill installed
We need to refactor 3 independent packages (packages/ppe-guard, packages/standby-mcp, packages/niche-miner) and update the root CI workflow. Plan and execute this safely with parallel agents.
- Read your context and instructions
- Compiled the multi-agent-worktree-orchestrator
Multi-Agent Worktree & Throughput Plan
| Stream | Execution Route | Ownership Scope (Disjoint) | Build Lock | Verification Receipt |
|---|---|---|---|---|
| Wave 1A (Main Agent) | MAIN_AGENT_DIRECT | packages/ppe-guard/** + shared tsconfig.base.json | Owns root build/test lock | .verify/wave1a-ppe.log (exit 0) |
| Wave 1B (Scout Subagent) | CONSOLIDATED_BATCH_SUBAGENT (Read-Only) | packages/standby-mcp/** + packages/niche-miner/** (34.2k tokens) | Zero lock touch | Returns compact file:line diff spec |
| Wave 2 (Worktree Writer) | .worktrees/agent-niche-miner | packages/niche-miner/** | Isolated verify -> cherry-pick | .verify/wave2-merge.log (exit 0) |
Connects securely to your tools. The creator never sees your data.
About this skill
Managing multiple AI agents across a large codebase often leads to git index contention, merge conflicts, and massive token waste. This skill implements a strict orchestration protocol using Git worktrees to run parallel subagents in isolated environments with zero file overlap.
What it does
- Route tasks efficiently by applying hard token thresholds to determine if a subagent is worth the bootstrap cost.
- Isolate workstreams using separate Git worktrees to prevent agents from overwriting each other or stalling on file locks.
- Enforce checkpointing with a mandatory 4-line pre-flight check to identify shared mutable state and blocking dependencies.
- Manage agent lifecycle by automating worktree creation, directory-level permission scoping, and cleanup.
How it works
- Analyze Task Shape by checking if the work involves disjoint files or exceeds the token threshold for the main context window.
- Execute Checkpoint to identify shared types or build caches that prevent parallel execution.
- Provision Worktree by spawning a branch in a hidden directory for the specific subagent task.
- Verify and Merge using a required verification receipt log before pruning the worktree and returning to the main branch.
Frameworks & tools
This skill is designed for Claude Code, Cursor, and Codex CLI. It utilizes Git worktrees and shell commands for environment isolation and filesystem management.
Why this beats prompting it yourself
Generic prompts often lead to agents hallucinating file paths or conflicting with their own edits in a shared directory. This skill provides a rigid logic gate that prevents $20+ token waste per subagent by enforcing economic thresholds and physical file isolation.
Use cases
- Monorepo migrations where multiple packages need simultaneous updates without shared build lock contention.
- High-volume refactoring across thousands of lines of code that would otherwise exceed context limits.
- Parallel feature implementation in disjoint modules where atomic rollbacks are required for each sub-task.
Known limitations
Does not support builds sharing a single locked directory like target/ or DerivedData/ in parallel. Max active subagents is capped at 2 to maintain stability.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Not suitable for tasks sharing a single build lock or directory. Designed for a maximum of two concurrent subagents. Requires Git worktree support.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 10 days ago
- Software
Git