hive-mcp

hive-mcp

Integrates Jira and Sourcegraph, enabling LLMs to retrieve and comment on Jira issues and perform code searches using Sourcegraph.

Category
访问服务器

README

Hive MCP Server

A Model Context Protocol (MCP) server that provides integrations for Jira and Sourcegraph. This allows LLMs to interact with Jira issues and search code across repositories using Sourcegraph.

Package name: hive-mcp
Tool prefix: Service name (e.g., jira_*, sourcegraph_*)
Full tool names in LLM: mcp__hive-mcp__jira_get_issue etc.

Features

Jira Integration

  • Get Issue: Retrieve detailed information about Jira issues
  • Download Attachments: Download attachments from Jira issues
  • Add Comments: Post comments to Jira issues

Sourcegraph Integration

  • Code Search: Search code across repositories with powerful query operators (GraphQL)
  • Deep Search: AI-powered conversational code search that understands natural language questions (REST API v6.7+)

Installation

Prerequisites

  • Node.js 18 or higher
  • Jira Server/Data Center instance with API access
  • Sourcegraph instance with API access
  • GitHub account with access to the @ocuco organization

Install from GitHub Packages

This package is published as a private package on GitHub Packages. To install it, you need to authenticate with GitHub first.

1. Create a GitHub Personal Access Token (PAT)

  1. Go to GitHub Settings > Developer Settings > Personal Access Tokens > Tokens (classic)
  2. Click "Generate new token" > "Generate new token (classic)"
  3. Give it a descriptive name (e.g., "NPM Package Access")
  4. Select the following scopes:
    • read:packages - Download packages from GitHub Package Registry
    • write:packages - Upload packages to GitHub Package Registry (only needed for publishing)
  5. Click "Generate token" and copy the token

2. Configure NPM to use GitHub Packages

Create or edit your ~/.npmrc file (in your home directory) and add:

@ocuco:registry=https://npm.pkg.github.com
//npm.pkg.github.com/:_authToken=YOUR_GITHUB_TOKEN

Replace YOUR_GITHUB_TOKEN with the token you created in step 1.

Note: On Windows, your home directory is %USERPROFILE% (usually C:\Users\YourUsername)

3. Install the package

npm install -g @ocuco/hive-mcp

Install from Source

git clone https://github.com/ocuco/hive-mcp.git
cd hive-mcp
npm install
npm run build
npm link

Configuration

Create a .env file in your project root or set environment variables:

# Jira Configuration (On-Premise Server/Data Center)
JIRA_BASE_URL=https://jira.your-company.com
JIRA_API_TOKEN=your_personal_access_token

# Sourcegraph Configuration
SOURCEGRAPH_URL=https://sourcegraph.company.com
SOURCEGRAPH_TOKEN=your_sourcegraph_token

Getting API Credentials

Jira Personal Access Token (PAT)

For Jira Server/Data Center on-premise installations:

  1. Log in to your Jira instance
  2. Navigate to your ProfilePersonal Access Tokens
  3. Click Create token
  4. Give it a name (e.g., "MCP Server Access")
  5. Set expiration date (optional, recommended for security)
  6. Click Create and copy the token
  7. Use this token as JIRA_API_TOKEN

Important Notes:

  • Personal Access Tokens are available in Jira 8.14 and later
  • For older versions, contact your Jira administrator for API access options
  • Keep your token secure and never commit it to version control
  • The token is used with Bearer authentication
  • Self-signed SSL certificates are automatically handled

Sourcegraph Token

  1. Go to your Sourcegraph instance
  2. Navigate to Settings > Access tokens
  3. Create a new access token
  4. Copy the token and use it as SOURCEGRAPH_TOKEN

Usage with Claude Desktop

Add to your Claude Desktop configuration file:

MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "hive-mcp": {
      "command": "node",
      "args": ["/path/to/hive-mcp/dist/index.js"],
      "env": {
        "JIRA_BASE_URL": "https://jira.your-company.com",
        "JIRA_API_TOKEN": "your_personal_access_token",
        "SOURCEGRAPH_URL": "https://sourcegraph.company.com",
        "SOURCEGRAPH_TOKEN": "your_sourcegraph_token"
      }
    }
  }
}

Or if installed globally from GitHub Packages:

{
  "mcpServers": {
    "hive-mcp": {
      "command": "hive-mcp",
      "env": {
        "JIRA_BASE_URL": "https://jira.your-company.com",
        "JIRA_API_TOKEN": "your_personal_access_token",
        "SOURCEGRAPH_URL": "https://sourcegraph.company.com",
        "SOURCEGRAPH_TOKEN": "your_sourcegraph_token"
      }
    }
  }
}

Note: After installing from GitHub Packages with npm install -g @ocuco/hive-mcp, the command name remains hive-mcp (defined in package.json bin field).

Available Tools & Resources

Tools

All tools are prefixed with their service name for easy identification in LLM tool lists.

Full name format in LLM: mcp__hive-mcp__<tool_name>

Jira Tools

jira_get_issue

Get detailed information about a Jira issue.

Parameters:

  • issueIdOrKey (string): The issue key (e.g., "PROJ-123") or ID

Example:

{
  "issueIdOrKey": "PROJ-123"
}

jira_get_comments

Get all comments from a Jira issue. Returns comment history with authors and timestamps.

Parameters:

  • issueIdOrKey (string): The issue key (e.g., "PROJ-123") or ID

Returns:

  • Total comment count
  • Array of comments with:
    • Author name and email
    • Created and updated timestamps
    • Comment body (Atlassian Document Format)

Example:

{
  "issueIdOrKey": "PROJ-123"
}

jira_add_comment

Add a comment to a Jira issue.

Parameters:

  • issueIdOrKey (string): The issue key (e.g., "PROJ-123") or ID
  • comment (string): The comment text to add

Sourcegraph Tools

sourcegraph_search_code

Search code across repositories.

Parameters:

  • query (string): Search query with optional operators
    • repo:owner/name - Filter by repository
    • file:path - Filter by file path
    • lang:python - Filter by language
    • And more...

Example:

{
  "query": "function handleRequest repo:myorg/myrepo lang:typescript"
}

sourcegraph_deep_search

AI-powered Deep Search using natural language questions. This feature creates a conversation with Sourcegraph's AI agent (requires v6.7+).

Parameters:

  • question (string): Natural language question about your codebase
  • timeout_seconds (number, optional): Max wait time for answer (default: 60)

Example:

{
  "question": "How does authentication work in this codebase?"
}

Response includes:

  • AI-generated answer
  • Relevant source code references
  • Suggested follow-up questions
  • Conversation ID for follow-ups
  • Shareable URL

Note: Deep Search is async and may take 10-60 seconds. It uses credits/quota on Sourcegraph Enterprise.


Resources

MCP Resources allow LLMs to read Jira attachments like local files. The server handles caching and download automatically.

Resource URI Format:

jira://issues/{issueKey}/attachments

List all attachments for a specific Jira issue.

Example:

jira://issues/PROJ-123/attachments

Returns:

  • List of all attachments with metadata
  • Each attachment includes its resource URI for direct access

jira://attachments/{attachmentId}

Read specific attachment content.

Example:

jira://attachments/10041

Behavior:

  • Images <500KB: Returns image blob (viewable directly)
  • Files <500KB: Returns base64 content
  • Files >500KB: Returns metadata + warning
  • Automatic caching: Small files cached for 5 minutes

Advantages:

  • ✅ LLM reads attachments like local files
  • ✅ No repeated downloads (smart caching)
  • ✅ No authentication errors (handled by MCP)
  • ✅ Automatic size management

Usage in Claude:

"Read jira://attachments/10041"
"Show me all attachments from PROJ-123"

Development

# Install dependencies
npm install

# Build
npm run build

# Development mode (watch for changes)
npm run dev

# Run the server
npm start

Publishing to GitHub Packages

Note: Only maintainers with write access to the @ocuco organization can publish.

Prerequisites for Publishing

  1. Create a GitHub Personal Access Token with write:packages scope (see Installation section)
  2. Ensure your local ~/.npmrc has the authentication configured
  3. Make sure you're logged in to npm: npm whoami --registry=https://npm.pkg.github.com

Publishing Steps

  1. Update version (choose one):

    npm version patch  # 1.0.0 -> 1.0.1
    npm version minor  # 1.0.0 -> 1.1.0
    npm version major  # 1.0.0 -> 2.0.0
    
  2. Build the project:

    npm run build
    
  3. Publish to GitHub Packages:

    npm publish
    
  4. Push changes and tags:

    git push
    git push --tags
    

Package Visibility

The package is set as private ("access": "restricted" in publishConfig). Only members of the @ocuco organization with proper permissions can:

  • View the package
  • Download and install the package
  • Publish new versions (with write access)

To manage package access:

  1. Go to the package page on GitHub
  2. Find the hive-mcp package
  3. Go to "Package settings" > "Manage Actions access" to configure permissions

Architecture

hive-mcp/
├── src/
│   ├── index.ts              # MCP server entry point
│   ├── config.ts             # Environment configuration
│   ├── types.ts              # Shared types
│   ├── jira/
│   │   ├── client.ts         # Jira REST API client
│   │   ├── tools.ts          # Jira MCP tools
│   │   └── types.ts          # Jira types
│   └── sourcegraph/
│       ├── client.ts         # Sourcegraph GraphQL client
│       ├── tools.ts          # Sourcegraph MCP tools
│       └── types.ts          # Sourcegraph types
└── dist/                     # Compiled output

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选