Jenkins MCP Server
Enables interaction with Jenkins instances to manage jobs, builds, and nodes through the Model Context Protocol. It provides capabilities to trigger builds, read console outputs, and search log sections using regular expressions.
README
Jenkins MCP Server
MCP (Model Context Protocol) server for interacting with Jenkins at jenkins.scylladb.com. Provides tools for listing jobs, viewing builds, reading console output, triggering builds, and more.
Setup
Prerequisites
npm install
Credentials
The server uses these environment variables for authentication:
| Variable | Required | Description |
|---|---|---|
JENKINS_USER |
Yes | Your Jenkins username |
JENKINS_TOKEN |
Yes | Your Jenkins API token |
JENKINS_URL |
No | Jenkins base URL (default: https://jenkins.scylladb.com) |
Step 1: Get your Jenkins username
Your Jenkins username is the same as your ScyllaDB LDAP / SSO username. You can verify it by logging in to https://jenkins.scylladb.com and checking the name shown in the top-right corner.
Step 2: Generate an API token
- Log in to https://jenkins.scylladb.com
- Click your name in the top-right corner → Configure (or go directly to
https://jenkins.scylladb.com/user/<your-username>/configure) - Scroll down to the API Token section
- Click Add new Token, give it a descriptive name (e.g.
claude-code-mcp), and click Generate - Copy the token immediately — it is shown only once
Step 3: Verify access
Test that your credentials work:
curl -u "your-username:your-api-token" https://jenkins.scylladb.com/api/json?tree=mode
You should see {"mode":"NORMAL"} (or similar). If you get a 401/403, double-check your username and token.
Adding to Claude Code
Add the server to your Claude Code MCP configuration. Edit (or create) the file ~/.claude/settings.json:
{
"mcpServers": {
"jenkins": {
"command": "node",
"args": ["/path/to/jenkins-mcp/server.js"],
"env": {
"JENKINS_USER": "your-username",
"JENKINS_TOKEN": "your-api-token"
}
}
}
}
Replace /path/to/jenkins-mcp/server.js with the absolute path to server.js, and fill in your credentials. JENKINS_URL defaults to https://jenkins.scylladb.com and can be omitted.
You can also configure it at the project level by creating a .mcp.json file in the project root (this format is shared with OpenAI Codex):
{
"mcpServers": {
"jenkins": {
"command": "node",
"args": ["./server.js"],
"env": {
"JENKINS_USER": "your-username",
"JENKINS_TOKEN": "your-api-token"
}
}
}
}
Adding to OpenAI Codex
Codex uses the same .mcp.json project-level configuration as Claude Code (see above). Place the .mcp.json file in the project root and Codex will pick it up automatically.
Running Standalone (for testing)
JENKINS_USER=your-username JENKINS_TOKEN=your-api-token node server.js
The server communicates over stdio using the MCP protocol — it is not meant to be used directly from a terminal, but this verifies it starts without errors.
Development
Linting
npm run lint
Uses ESLint 9 with flat config (eslint.config.js).
CI/CD
GitHub Actions runs on every push and PR to main:
- Lint —
npx eslint .with Node 20
Driver Tests Overview
The Jenkins folder scylla-master/driver-tests contains CI jobs that test ScyllaDB compatibility with various CQL client drivers.
How Driver Tests Work
Each driver test job:
- Spins up a ScyllaDB cluster (typically via CCM)
- Runs the driver's test suite against it across multiple driver versions and protocol versions
- Reports results to Argus
The test infrastructure for each driver lives in a driver-matrix repo. These repos contain the configuration (which driver versions to test, which tests to run, version-specific workarounds, etc.). The actual driver source code lives in separate driver repos — ScyllaDB forks of upstream drivers (or ScyllaDB-native drivers in the case of Rust).
Job → Repository Mapping
Where to File Issues
- Test infrastructure issues (flaky tests, missing versions, matrix config problems) → file in the driver-matrix repo
- Driver bugs (incorrect behavior, crashes, protocol issues) → file in the driver repo
- ScyllaDB server bugs (discovered via driver tests) → file in scylladb/scylladb
Available MCP Tools
| Tool | Description |
|---|---|
list_jobs |
List jobs in a folder |
get_job_info |
Get detailed job info including parameters |
get_build_info |
Get info about a specific build |
get_console_output |
Get build console output (with optional tail) |
get_build_log_section |
Search console output with regex and context |
list_builds |
List recent builds for a job |
trigger_build |
Trigger a new build |
stop_build |
Abort a running build |
get_queue |
View the build queue |
list_views |
List Jenkins views |
get_view_jobs |
List jobs in a view |
search_jobs |
Search for jobs by name |
get_nodes |
List build nodes/agents |
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。