Claude DevFleet Multi-Agent Orchestration
Dispatch parallel Claude Code agents in isolated git worktrees with Muse: plan a project into a mission DAG, auto-dispatch on dependency completion, monitor progress, read structured reports.
- What
- Dispatch parallel Claude Code agents in isolated git worktrees with Muse: plan a project into a mission DAG, auto-dispatch on dependency completion, monitor progress, read structured reports.
- Cost
- Free
- Needs
- the Claude Code CLI installed and authenticated with an Anthropic API key — DevFleet plans and dispatches full Claude sessions, so it cannot plan or run anything without that, and each parallel agent multiplies token costs. Python 3.11+ for the backend server, and git. Without the CLI and key, the files below still install, but the connection step and any real use will stop there.
- Level
- Intermediate — the Claude Code CLI installed and authenticated…
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor: a full multi-agent orchestration workflow on top of the Claude DevFleet server. The flow is Plan, Dispatch, Monitor, Report. plan_project turns a plain-language description into a project with a mission DAG, chained by depends_on with auto_dispatch set; you show the plan, get approval, then dispatch the root mission while the rest auto-start as dependencies complete. Each agent runs in an isolated git worktree with full tooling and auto-merges on completion. Tools cover the whole lifecycle: create_project, create_mission, dispatch_mission, cancel_mission, wait_for_mission (blocks up to the timeout, prefer polling get_mission_status every 30-60 seconds for long runs), get_report (files changed, what was done, errors, next steps), get_dashboard (running agents, stats, activity), list_projects, list_missions. Concurrency runs up to 3 agents by default via DEVFLEET_MAX_AGENTS, with the mission watcher queueing the rest. Guidelines are firm: always confirm the plan with the user before dispatching unless they said go ahead, read a failed mission's report before retrying, never create circular dependencies, and merge conflicts stay on the agent's worktree branch for manual resolution. By @affaan-m, listed here with credit to its creator. From the affaan-m/ECC repository (MIT). Honest caveats: the DevFleet server is a separate project (github.com/LEC-AI/claude-devfleet) you install and run yourself, then connect over MCP to localhost:18801; verify the process on that port is the DevFleet binary you installed before connecting. Every agent is a full session, so token costs multiply; poll instead of blocking waits on long missions. Skill Harbor never reviews the code, review it yourself before use.
Version:
Quick answer
Claude DevFleet Multi-Agent Orchestration is a free Muse skill that dispatches parallel Claude Code agents in isolated git worktrees with Muse. The flow is Plan, Dispatch, Monitor, Report: turn a description into a mission DAG, show the plan, get approval, dispatch the root mission while the rest auto-start, then monitor progress and read reports.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit 4eb71d9Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites: the Claude Code CLI installed and authenticated with an Anthropic API key — DevFleet plans and dispatches full Claude sessions, so it cannot plan or run anything without that, and each parallel agent multiplies token costs. Python 3.11+ for the backend server, and git. Without the CLI and key, the files below still install, but the connection step and any real use will stop there. Install "Claude DevFleet Multi-Agent Orchestration" for me. It plans a project into a mission DAG, dispatches parallel Claude agents in isolated git worktrees through a local DevFleet server, and tracks their structured reports. Repository: https://raw.githubusercontent.com/affaan-m/ECC/4eb71d92a39cab44ad40ac9d8a6a5ccb4029d6c2/skills/claude-devfleet/SKILL.md Pinned version: install the code at commit 4eb71d92a39cab44ad40ac9d8a6a5ccb4029d6c2 exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Fetch into a temporary folder and look inside before doing anything else: the SKILL.md from the repository URL above, and a clone of the DevFleet server repo, https://github.com/LEC-AI/claude-devfleet (the skill drives that server, it is not included in the SKILL.md). 2. Integrity check: Skill Harbor reviewed this listing on 2026-10-10 and recorded the fingerprint below. I want you to compute the SHA-256 of the downloaded SKILL.md and compare: - SKILL.md: SHA-256 5fe6efbb5663699ba8e348b2a77fbf7c396ed1903bb95fa0c05c1174f81fabfc If the fingerprint does not match, I want you to stop before installing and tell me the files changed since Skill Harbor's review, so the new version is unverified. Show me your safety findings and give me three choices: (a) wait for Skill Harbor to re-review the listing (recommended), (b) analyze the new files in detail first and then decide, (c) install anyway knowing the risk. Then wait for my answer before doing anything else. 3. Safety check: I want to see your findings explicitly, in short separate items: - every network address the code contacts (expect: the backend listens on localhost port 18801 only; the agents it dispatches contact Anthropic's API using my configured key) - every shell command it runs (expect: a Python venv with pip install of backend/requirements.txt, and the uvicorn server start) - every place credentials, API keys or personal data are read, stored or sent (expect: my Anthropic key stays in my Claude CLI configuration; the backend keeps its data in a local database file) - any obfuscated or encoded code (expect: none) If anything looks suspicious, stop and explain it to me before continuing. One trust note: connecting a localhost MCP server is a trust decision, so before connecting, confirm the process listening on port 18801 is the DevFleet server you just installed. 4. Confirmation: before installing, give me a short summary of what the skill does plus your safety findings, and wait for my explicit "yes, install". 5. Install it: copy SKILL.md into the agent's skills directory, in a folder named "claude-devfleet". Then set up the backend from the clone, following its README's manual local setup: create a venv inside the clone and install backend/requirements.txt into it. Start the server from the backend folder (uvicorn app:app on port 18801), confirm it answers locally, then connect it: claude mcp add devfleet --transport http http://localhost:18801/mcp. 6. Verify with no network calls: frontmatter valid, skill name is claude-devfleet, the file is in place, and the backend venv exists with its requirements installed. 7. Report what was installed, where, and what I still need to do myself: keep the server running while I use DevFleet, set DEVFLEET_MAX_AGENTS to match my budget, and approve each mission plan before anything is dispatched. GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. Never ask me to paste secrets in chat: credentials go through the secure vault or environment variables. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
- It can run commands on your Muse’s computer.
- It can download and install software packages.
- It works with an API key or a token that you provide.
What the scan found — excerpts from the install prompt
- runs commands — - chaque commande shell qu'il exécute (attendu : un venv Python avec pip install de backend/requirements.txt,
- installs packages — - every shell command it runs (expect: a Python venv with pip install of backend/requirements.txt, and the uvi · - chaque commande shell qu'il exécute (attendu : un venv Python avec pip install de backend/requirements.txt,
- uses API keys or tokens — Prerequisites: the Claude Code CLI installed and authenticated with an Anthropic API key — DevFleet plans and · Prérequis : le CLI Claude Code installé et authentifié avec une clé API Anthropic — DevFleet planifie et envoi
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- 4eb71d9 · 2026-10-13
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026
- Creator identity
- Not verified
- Fingerprint watch
- Last checked · Nov 16, 2026 : no change detected in the files linked by this install prompt
- Install prompt analysis
- 2026-10-13 : runs commands · installs packages · uses API keys or tokens
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.