About Release Intel
Release Intel is an MCP server for developers preparing release notes from repository history. It compares Git refs and structures commits, merged pull requests, linked issues, changed files, and contributors as release evidence.
Key features
- 01Compare commits between Git refs
- 02Enrich commits with pull request and issue data
- 03Group merged pull requests by category
- 04Create structured release evidence reports
- 05Generate Markdown or JSON reports
Use cases
- 01Collect evidence for a version release
- 02Separate merged pull requests from direct commits
- 03Attach a release evidence report to a workflow summary
MCP support
Exposes tools to compare refs, retrieve categorized merged pull requests, and build a release summary. It also includes a CLI report command and a GitHub Action for workflow summaries.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y @barissozudogru/release-intel-mcpGITHUB_TOKENSecretGitHub token with repository read access, used for the GitHub API.Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
3Saved tool list · updated 2w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| base | string | Yes | Base ref (tag, branch, or commit SHA), the older point |
| head | string | Yes | Head ref (tag, branch, or commit SHA), the newer point |
| repo | string | Yes | GitHub repository name |
| owner | string | Yes | GitHub repository owner (user or organization) |
Details
From the project README
release-intel-mcp
Build release context from repository evidence instead of a blank prompt. The server compares two Git refs and correlates commits, merged pull requests, linked issues, labels, changed files, and contributors into a structured release record.
Tool page · npm · Listed in the official MCP Registry
Start in one minute
Create a fine-grained GitHub token with read-only access to the repositories you want to inspect. Repository contents and pull requests are sufficient for the current tools. No write permission is needed.
{
"mcpServers": {
"release-intel": {
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "your_read_only_token"
}
}
}
}
Then ask the client for a bounded range:
Use release-intel to build release evidence for owner/repository
from v1.4.0 to v1.5.0. Separate merged PRs from direct commits
and preserve links to the repository record.
The token stays in the local server process and is sent only to the GitHub API.
Proof on this repository
A verified run against the stable range v0.5.1...8a2f350 returned:
Repository: barissozudogru/release-intel-mcp
Commits: 14
Merged PRs: 1
Files changed: 12
Lines added: 286
Lines deleted: 135
Contributors: 1
The response also includes commit SHAs, messages, authors, pull request categories, linked issues, and warnings when GitHub truncates or fails part of the evidence lookup.
If this saves you time, consider starring the repository. It helps other developers find it.
Tools
| Tool | What it returns |
|---|---|
get_changes_between_refs | Commits enriched with merged pull request metadata and linked issues |
get_pull_requests_in_range | Merged pull requests grouped as breaking, feature, fix, docs, chore, dependency, release, or other |
get_release_summary | A complete release evidence object with aggregate statistics and contributors |
The categorization uses labels and conventional title signals. Repository labels, issue links, and pull request hygiene directly affect the result. The server provides evidence and structure; the final release narrative still needs review.
Release report command
The package includes a terminal command that produces Markdown by default:
GITHUB_TOKEN=your_read_only_token \
npx --yes --package @barissozudogru/release-intel-mcp \
release-intel-report owner/repository v1.4.0 v1.5.0
Use --json when another tool will consume the report:
release-intel-report owner/repository v1.4.0 v1.5.0 --json
GitHub Action
Attach the evidence report to the workflow summary before publishing a release:
name: Release evidence
on:
workflow_dispatch:
inputs:
from_ref:
required: true
to_ref:
required: true
jobs:
evidence:
runs-on: ubuntu-latest
permissions:
contents: read
pull-requests: read
steps:
- uses: barissozudogru/[email protected]
with:
from-ref: ${{ inputs.from_ref }}
to-ref: ${{ inputs.to_ref }}
token: ${{ secrets.GITHUB_TOKEN }}
This action reads repository data and writes Markdown to the job summary. It does not publish a release or modify repository content.
Client setup
Claude Desktop, Cursor, Windsurf, Cline, and similar clients
Use the stdio configuration from the quickstart. The server entry is identical across these clients even when their config file locations differ.
VS Code with Copilot
Create .vscode/mcp.json:
{
"servers": {
"release-intel": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@barissozudogru/release-intel-mcp"],
"env": {
"GITHUB_TOKEN": "your_read_only_token"
}
}
}
}
Streamable HTTP
GITHUB_TOKEN=your_read_only_token \
npx @barissozudogru/release-intel-mcp --http
The MCP endpoint is http://localhost:3000/mcp and the health endpoint is http://localhost:3000/health. Set PORT to change the port.
Docker
docker build -t release-intel-mcp .
docker run -p 3000:3000 \
-e GITHUB_TOKEN=your_read_only_token \
release-intel-mcp
Connect an HTTP client to http://localhost:3000/mcp.
Local development
npm install
npm test
npm run build
GITHUB_TOKEN=your_read_only_token node dist/index.js
Requirements: Node.js 18 or newer.
See CONTRIBUTING.md for the development workflow.
Security and limits
- Use a fine-grained, read-only token restricted to the required repositories.
- The server never needs permission to write releases, issues, or pull requests.
- GitHub's compare endpoint limits very large ranges. The response reports truncation.
- Unlinked direct commits remain visible but cannot inherit pull request metadata.
- Release categories are evidence-backed heuristics, not editorial conclusions.
License
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.