M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
C

Claw Orchestrator

Community
Unified runtime to orchestrate Claude Code, Codex, Gemini, and other coding CLIs with multi-agent collaboration.
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #272 of 438
Stars
★ 589 Very popular
Transport
stdio (local process)
Runtime
Node.js
Credentials
No credential needed
License
MIT
Last commit
Tools
1
49FMRS · D

Claw Orchestrator is a powerful third-party MCP server that unifies multiple coding agents through a single interface, offering advanced features like persistent sessions, multi-agent councils, and autonomous loops. It is well-suited for professional users requiring complex agent orchestration, but its non-official nature and potential security risks should be considered, and it should be used in a controlled environment.

Strongest · Documentation 13/20 Weakest · Reliability 6/20

Reliability
6/20
Security and permissions
8/20
Maintenance
10/20
Documentation
13/20
Setup experience
12/20
Why each score
Reliability 6/20
The README presents a coherent architecture, a quick-start flow, and an engine compatibility matrix with tested versions, so the happy path is plausible and a test command is referenced in the contributing guide. However, the server manifest was not cached and the supplied evidence contains no CI workflow contents or committed tests, so the MCP init/tool-list handshake and the claimed 65 tools cannot be verified against real behavior. Under static calibration the README is untrusted evidence; with no verifiable execution evidence the score must stay below 12.
Security and permissions 8/20
No red-line findings: no malware indicators, credential theft, covert exfiltration, or real tokens in install examples are visible in the source. The README documents BYOK, cookie-based dashboard auth, and a CLAWO_MCP_TOOLS allowlist, showing some boundary control. Yet the server's core function is executing coding CLIs as subprocesses plus autonomous and deploy-style workflows, and the README gives no confirmation mechanism for dangerous operations, no credential-scoping details, and no data-flow disclosure, so least privilege and explicit confirmation remain incomplete.
Maintenance 10/20
The evidence indicates an active project: the repository is not archived, and the README shows an MIT license, an npm version badge, a CI badge, a contributing guide with build/lint/format/test checks, and a defined update path. Deductions: the badges and repo stats are untrusted discovery signals, and the supplied source gives no verifiable commit frequency, release cadence, dependency update history, issue-response evidence, or security-response channel, leaving governance and versioning gaps.
Documentation 13/20
The README is layered and detailed: a feature table with per-capability doc references, quick-start examples, CLI/OpenClaw/MCP integration commands, an engine compatibility matrix with tested versions, and mention of per-host stdio config snippets and a tool allowlist. Deductions: the actual tool-parameter list, limits/cost notes, and troubleshooting guidance exist only as references and are not present in the evidence, and authentication details are thin, so the docs cite more than they show.
Setup experience 12/20
The setup path is clear in principle: one global npm install puts clawo-mcp on PATH, then register it in a supported MCP client, with the README claiming per-host stdio config snippets for mainstream clients such as Claude Desktop, Cursor, Cline, and Windsurf. Deductions: the actual config snippets are not in the evidence, each coding engine CLI must be pre-installed separately, and Node/platform prerequisites are unspecified; with no verifiable CI/test evidence the static calibration caps setup at 15, hence 12.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesWrites / deletes local filesRuns commands or codeUses the network

Best for

  • Developers and teams needing a unified interface to manage multiple coding agents.
  • Users who want to automate multi-agent workflows such as plan-code-review loops.
  • Individuals who wish to integrate MCP into their existing agent workflows.
  • Those interested in exploring agent orchestration or persistent session management.

Not for

  • Simple use cases that do not require multi-agent orchestration or only use a single coding CLI.
  • Beginners preferring a full GUI without relying on external CLIs.
  • Environments that are conservative about non-official third-party tools.

Required permissions

  • This server executes locally installed coding CLIs (e.g., `claude`, `codex`, `agy`, `agent`, `opencode`) as subprocesses.
  • It requires read/write access to working directories and git worktrees.
  • It may open network ports (default 18796) to serve a dashboard and HTTP API.
  • It may use the file system to store session state and configuration.

Risks and side effects

  • As a third-party package, it is not officially certified by the upstream products, posing potential security risks.
  • Executing user-supplied agent commands can lead to arbitrary code execution; use in a controlled environment.
  • Dependence on external CLI versions may cause functionality issues due to version incompatibilities.
  • Network exposure (e.g., port 18796) could lead to unauthorized access; configure authentication if exposed.

Setup

Before you start

Runtime：Node.js

Other optional settings (1)
CLAWO_MCP_TOOLS optional An allowlist limiting which tools the MCP server exposes for tight tool budgets; optional, documented in mcp.md.
  1. Install the npm package globally: npm install -g @enderfga/claw-orchestrator (this puts clawo-mcp on PATH).
  2. Register clawo-mcp as a stdio MCP server in your MCP-compatible host (e.g., Claude Desktop, Cursor, Zed, Windsurf, Continue, Cline).
  3. Optionally set CLAWO_MCP_TOOLS environment variable to restrict the set of available tools.

Check that it works

After npm install -g @enderfga/claw-orchestrator, confirm clawo-mcp is on PATH and register it with your MCP client; the clawo-mcp tool appearing in the client's tool list proves the connection.

Troubleshooting

  1. Ensure all required CLIs (e.g., `claude`, `codex`) are installed and on PATH.
  2. Verify that `clawo-mcp` is correctly configured as a stdio MCP server in your client.
  3. Check whether `CLAWO_MCP_TOOLS` environment variable includes the desired tool allowlist.
  4. Inspect CLI command output (e.g., `clawo serve`) logs for error messages.
  5. Confirm firewall allows port 18796 if using the dashboard.

Things to try

Once connected, you can ask your AI assistant things like:

  • Start a session named fix-tests with the claude engine and have it fix the failing tests
  • Fan out the same task to claude and codex engines in parallel and synthesize their answers
  • Show me the status of all sessions currently running in the clawo dashboard

Tools 1

clawo-mcp writes
Provides the MCP interface to Claw Orchestrator, exposing its 65 tools to MCP-compatible clients.

Use cases

Invoke coding CLIs like Claude Code, Codex, or Cursor Agent as persistent sessions directly from an MCP client.
Create multi-agent councils where parallel agents in isolated git worktrees vote on consensus.
Run autonomous Planner/Coder/Reviewer loops for workspace iteration.
Generate and deploy a web app from a five-question interview (Ultraapp mode).
Integrate multiple coding engines under one interface for unified tool orchestration.

Supported clients

Claude Desktop
Cursor
Zed
Windsurf
Continue
Cline

Listed from the project's documentation, not tested by this site.

Overview

Claw Orchestrator is a runtime that wraps Claude Code, Codex, Antigravity, Cursor Agent, OpenCode, or any custom CLI into persistent, programmable sessions. It enables multi-agent councils, autonomous Planner/Coder/Reviewer loops, and even generates deployed web apps from a five-question interview. It exposes a 65-tool API accessible via CLI, OpenClaw gateway, Model Context Protocol, or directly from TypeScript, with an embedded dashboard.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 5221a89ffa66 Data synced 2026-10-12 Read the FMRS scoring method