release-intel
release-intel-mcp
一个从 GitHub 仓库数据生成发布情报的 MCP 服务器。它关联任意两个 git ref 之间的提交、拉取请求、问题与贡献者,并返回结构化上下文,供 AI 综合生成发布说明、变更日志和发布摘要。
工具
get_changes_between_refs
获取两个 git ref 之间的所有提交,并附带相关的 PR 元数据、作者信息和关联问题。
字段 | 类型 | 描述 |
| string | GitHub 仓库所有者或组织 |
| string | GitHub 仓库名称 |
| string | 基础 ref(较早的标签、分支或 SHA) |
| string | 目标 ref(较新的标签、分支或 SHA) |
get_pull_requests_in_range
获取两个 ref 之间所有已合并的 PR,并按标签自动分类为:breaking、feature、fix、docs、chore、dependencies、other。
输入字段与 get_changes_between_refs 相同。
get_release_summary
生成一个结构化的发布上下文对象,包含提交数据、PR 元数据、关联问题、贡献者列表和汇总统计。
字段 | 类型 | 描述 |
| string | GitHub 仓库所有者 |
| string | GitHub 仓库名称 |
| string | 上一个发布标签(基础) |
| string | 新的发布标签或 HEAD |
Related MCP server: MCP Releases Server
设置
所有选项都需要具有 repo 读取权限的 GitHub 个人访问令牌。
在此处生成:https://github.com/settings/tokens
选项 A:stdio(本地进程)
标准方式:MCP 客户端将服务器作为本地子进程启动。
Claude Desktop
配置文件:~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"release-intel": {
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}Claude Code
claude mcp add release-intel -e GITHUB_TOKEN=ghp_your_token -- npx -y @barissozudogru/release-intel-mcpCursor
配置文件:~/.cursor/mcp.json
{
"mcpServers": {
"release-intel": {
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}Windsurf
配置文件:~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"release-intel": {
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}VS Code + Copilot
配置文件:.vscode/mcp.json
{
"servers": {
"release-intel": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}Cline
配置文件:~/.cline/mcp_settings.json(或通过 Cline 扩展设置界面)
{
"mcpServers": {
"release-intel": {
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}Continue.dev
配置文件:~/.continue/config.yaml
mcpServers:
- name: release-intel
command: npx
args:
- -y
- "@barissozudogru/release-intel-mcp"
env:
GITHUB_TOKEN: ghp_your_tokenZed
配置文件:~/.config/zed/settings.json
{
"context_servers": {
"release-intel": {
"command": {
"path": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}
}JetBrains (AI Assistant plugin)
{
"mcpServers": {
"release-intel": {
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "ghp_your_token"
}
}
}
}选项 B:HTTP(远程 / 无状态)
将服务器作为 HTTP 端点运行。适用于远程客户端、团队共享部署,或偏好基于 URL 连接的客户端。
GITHUB_TOKEN=ghp_your_token npx @barissozudogru/release-intel-mcp --http默认在 3000 端口启动。设置 PORT 可更改端口。
Cursor (HTTP)
{
"mcpServers": {
"release-intel": {
"url": "http://localhost:3000/mcp"
}
}
}VS Code + Copilot (HTTP)
{
"servers": {
"release-intel": {
"type": "http",
"url": "http://localhost:3000/mcp"
}
}
}Windsurf (HTTP)
{
"mcpServers": {
"release-intel": {
"serverUrl": "http://localhost:3000/mcp"
}
}
}Continue.dev (HTTP)
mcpServers:
- name: release-intel
type: streamable-http
url: http://localhost:3000/mcp健康检查端点位于 GET /health。
选项 C:Docker
docker build -t release-intel-mcp .
docker run -p 3000:3000 -e GITHUB_TOKEN=ghp_your_token release-intel-mcp容器默认以 HTTP 模式启动。MCP 端点为 http://localhost:3000/mcp。
环境变量
变量 | 必需 | 默认值 | 描述 |
| 是 | - | GitHub 个人访问令牌(repo 范围) |
| 否 | stdio | 设置为 |
| 否 | 3000 | HTTP 端口(仅 HTTP 模式) |
本地开发
git clone https://github.com/barissozudogru/release-intel-mcp.git
cd release-intel-mcp
npm install
npm run build
GITHUB_TOKEN=ghp_... node dist/index.js许可证
MIT
Available Tools
3 toolsget_changes_between_refsGet Changes Between RefsB
Get all commits between two git refs enriched with associated PR metadata, author information, and linked issues. Uses the GitHub compare API.
Inputs: base (required), head (required), repo (required), owner (required).
get_pull_requests_in_rangeGet Pull Requests in RangeA
Get all merged pull requests between two refs with full metadata including labels, linked issues, review counts, and files changed. PRs are categorized by label into: breaking, feature, fix, docs, chore, dependencies, other.
Inputs: base (required), head (required), repo (required), owner (required).
get_release_summaryGet Release SummaryA
Generate a structured release context object ready for AI synthesis into release notes. Combines commit data, PR metadata, linked issues, contributor list, and aggregate statistics for the range between two tags.
Inputs: repo (required), owner (required), to_tag (required), from_tag (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 3 tools
The three tools are largely distinct: one fetches commits, one fetches PRs, and one combines both into a summary. However, get_changes_between_refs and get_pull_requests_in_range both retrieve PR-related data, which could cause slight confusion for an agent deciding between them. The descriptions provide enough clarity to mitigate overlap.
All tool names follow a consistent 'get_' verb prefix followed by descriptive noun phrases in snake_case (changes_between_refs, pull_requests_in_range, release_summary). This pattern is predictable and aligns with common conventions.
With only 3 tools, the server is lean but each tool serves a clear purpose in the release-intel workflow: retrieving commits, retrieving PRs, and generating a summary. The count is minimal yet sufficient for the stated domain, though it sits at the low end of the ideal range.
The server covers the core operations for generating release context: commits, PRs, and a combined summary. However, there is no tool to list available tags/refs, which would be a natural precursor to using the other tools, and no way to directly fetch issues outside of PRs. These gaps could require agents to work around them.
Maintenance
Related MCP Connectors
A MCP server built for developers enabling Git based project management with project and personal…
Create, deploy, and operate MCP servers directly from your GitHub repositories.
One MCP server for all your MCP servers: code mode, per-tool permissions, approvals and audit.
Repository knowledge graph MCP server for codebase understanding and debugging.
Related MCP Servers
- AlicenseAqualityBmaintenanceA MCP server that bridges LLMs with GitHub repository management, enabling automated analysis of pull requests, issue management, tag creation, and release management through natural language.446Apache 2.0
- FlicenseNot gradedqualityDmaintenanceAn MCP server that provides GitHub release information through the Model Context Protocol. It features a tool to retrieve the latest stable release version of any specified GitHub repository.-
- AlicenseNot gradedqualityDmaintenanceA production-ready MCP server that provides AI assistants with comprehensive GitHub developer tooling including PR analysis, code review, changelog generation, dependency auditing, commit summarization, and refactoring suggestions.MIT
- AlicenseAqualityDmaintenanceMCP server for developer workflow automation — smart commits, secret scanning, PR descriptions, and more.5MIT