Registry / devops / codeloop-mcp-server

codeloop-mcp-server

library0.1.73jsnpmunverified

CodeLoop MCP server (v0.1.73) provides 29 tools for AI coding agents to verify, self-repair, and gate-check code changes. It runs locally, never spawns its own LLM calls—orchestrates real test runners, screenshot/video tools, and gate evaluators—posting structured artifacts back into the agent chat. Zero extra LLM tokens, zero rate-limit risk. Works on macOS, Linux, Windows. Usually installed via the codeloop CLI.

devops testing

npm install codeloop-mcp-server

github ↗package ↗homepage ↗

INSTALL

IMPORT

SIG · CODELOOP-MCP-SERVE

codeloop-mcp-server

devopsjavascriptv0.1.73

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

Server

✓ import { Server } from '@modelcontextprotocol/sdk/server/index.js'

✗ import { Server } from 'codeloop-mcp-server'

codeloop-mcp-server is meant to be run as a CLI, not imported programmatically. Use the MCP SDK directly to build custom servers.

codeloop_verify (tool)

✓ Via MCP client: call tool 'codeloop_verify' with arguments

✗ const verify = require('codeloop-mcp-server').verifyTool

Tools are exposed over MCP transport, not as JS exports. Use StdioClientTransport to invoke.

default export / CLI

✓ npx -y codeloop-mcp-server

✗ npm install codeloop-mcp-server && require('codeloop-mcp-server')

The package is a CLI binary; programmatic usage is not supported. Use the CLI or codeloop wrapper.

Shows how to add the CodeLoop MCP server to Cursor or Claude Code via mcp.json, and how to instruct the agent to use the verify-diagnose-fix-gate loop.

// 1. Add to your MCP client config (e.g., Cursor mcp.json): { "mcpServers": { "codeloop": { "command": "npx", "args": ["-y", "codeloop-mcp-server"], "env": { "CODELOOP_API_KEY": process.env.CODELOOP_API_KEY ?? '' } } } } // 2. Instruct your AI agent: // After every code change, call codeloop_verify. If it fails, call codeloop_diagnose, fix, then re-verify. Do not declare the task done until codeloop_gate_check returns ready_for_review with confidence >= 94%.

Debug

Known issues

breakingcodeloop-mcp-server is a CLI tool, not a library. Attempting to import it programmatically will fail.

fix

Use npx -y codeloop-mcp-server in your MCP client configuration. Do not require/import.

affects: >=0.1.0

gotchaAll tool names are prefixed with 'codeloop_'. Omitting the prefix will result in 'Tool not found' errors.

fix

Always call tools by their full name (e.g., codeloop_verify, not verify).

affects: >=0.1.0

deprecatedThe env var CODELOOP_KEY was renamed to CODELOOP_API_KEY in v0.1.60. Using the old name will cause authentication failures.

fix

Update your config to use CODELOOP_API_KEY.

affects: <0.1.60

gotchaRunning codeloop-mcp-server without CODELOOP_API_KEY set will cause all API-dependent tools (e.g., codeloop_release_readiness) to fail silently or return errors.

fix

Set the CODELOOP_API_KEY environment variable to your CodeLoop API key.

affects: >=0.1.0

Errors

Common errors & fixes

Tool "verify" not found. Available tools: codeloop_verify, codeloop_diagnose, ...

Calling tools without the 'codeloop_' prefix.

fix

Use the full tool name: codeloop_verify.

Error: CODELOOP_API_KEY is not set

Missing API key environment variable.

fix

Set CODELOOP_API_KEY in your environment or mcp.json env block.

Cannot find module 'codeloop-mcp-server'

Trying to require() the package programmatically instead of running as CLI.

fix

Use npx -y codeloop-mcp-server or install the codeloop CLI wrapper.

Error: spawn npx ENOENT

Node.js not installed or not in PATH when using npx.

fix

Install Node.js >=18 and ensure it's in your PATH. Alternatively, install globally via npm install -g codeloop-mcp-server.

Upgrade

Version history

0.1.73latest on npm

Audit

Dependencies

@modelcontextprotocol/sdkrequiredMCP server SDK for tool definition and transport

Agent activity

14 hits · last 30 days

node

OpenAI (training)

Resources

githubhttps://github.com/OrienjoJi/codeloop-internal ↗

packagecodeloop-mcp-server ↗

homepagehttps://codeloop.tech ↗