CodeReview.MCP 0.7.0

<p align="center"> <img src="REBUSS.Pure.png" alt="REBUSS.Pure" /> </p>

🚀 REBUSS.Pure – AI Code Review That Focuses Only on What Matters

Stop sending irrelevant code to AI.
Send only the right context — and understand Pull Requests faster.

💡 What is this?

REBUSS.Pure is a lightweight MCP server that enables AI agents (GitHub Copilot, ChatGPT, Claude) to perform high-signal code reviews by providing only the context that actually matters.

Instead of overwhelming the model with your entire repository, REBUSS.Pure:

• 🔍 analyzes Azure DevOps and GitHub Pull Requests
• 📄 provides only relevant code changes
• 🧠 enables focused code review & self-review
• ⚡ delivers minimal, high-signal context

🎯 Why this exists

Most AI workflows today:

• ❌ send too much code
• ❌ drown the model in noise
• ❌ produce generic, low-quality feedback

REBUSS.Pure changes the approach:

• ✅ sends only relevant context
• ✅ reduces noise, not just tokens
• ✅ helps AI focus on what actually matters

👉 built for real-world code review, not demos

🧠 Core idea

AI doesn’t need more code.
It needs the right code.

Instead of:

❌ full repo → LLM

You get:

LLM → MCP → high-signal context only

✨ Key Features

• 🔹 Azure DevOps and GitHub Pull Request integration
• 🔹 High-signal, diff-based AI context
• 🔹 Local self-review (no network required)
• 🔹 No repo cloning needed
• 🔹 Incremental, on-demand data access
• 🔹 Ready-to-use review prompts
• 🔹 Works with any MCP-compatible agent
• 🔹 Authentication via Azure CLI, GitHub CLI (gh auth), or PAT
• 🔹 Auto-detects VS Code and Visual Studio
• 🔹 Auto-detects provider from Git remote URL

🔒 Security & Privacy

Your source code never leaves your machine.

REBUSS.Pure runs as a local process on your workstation. It does not upload, store, or relay your code to any external service. The MCP server acts as a controlled gateway between your AI agent and the data it actually needs:

• Local processing only — the server runs on localhost; no outbound code transmission occurs.
• Minimal data exposure — the AI model receives only relevant context, not the full repository.
• Azure DevOps stays yours — when fetching PR data, requests go directly to your organization's Azure DevOps APIs using your credentials. No intermediary services are involved.
• GitHub stays yours — GitHub API requests go directly to api.github.com using your personal access token. No intermediary services are involved.
• Offline self-review — local review (#self-review) operates entirely without network access. Git operations run against your local repository; nothing is sent anywhere.
• No telemetry, no tracking — the server collects zero usage data and phones home to nobody.

In short: REBUSS.Pure gives AI agents precise, scoped access to exactly the context they need — and nothing more.

🆚 Compared to typical AI workflows

⚡ Quick Start

1. Install

Option A — .NET global tool (recommended)

dotnet tool install -g CodeReview.MCP

Option B - PowerShell

irm https://raw.githubusercontent.com/rebuss/CodeReview.MCP/master/install.ps1 | iex

Option C - Bash

curl -fsSL https://raw.githubusercontent.com/rebuss/CodeReview.MCP/master/install.sh | bash

2. Initialize in your repo

cd /path/to/your/repo
rebuss-pure init

This will:

• ✔ detect your IDE (VS Code → .vscode/mcp.json, Visual Studio → .vs/mcp.json)
• ✔ generate MCP server configuration
• ✔ copy review prompts to .github/prompts/
• ✔ copy instruction files to .github/instructions/ (for GitHub Copilot custom instructions)
• ✔ authenticate via Azure CLI (opens browser for login) or accept a GitHub PAT

Global mode (-g)

If Visual Studio does not detect the local .vs/mcp.json file in your repository, use the global flag:

cd /path/to/your/repo
rebuss-pure init -g

This writes the MCP configuration to the user-level paths (~/.vs/mcp.json and ~/.vscode/mcp.json) instead of the repository-local directories. The global config points --repo to the current repository, so it works for any workspace you open.

Switching between repositories: If you use multiple repositories, run rebuss-pure init -g in the target repository before switching to it. This updates the global configuration to point to the correct workspace.

After updating the tool: If you run dotnet tool update -g CodeReview.MCP, run rebuss-pure init again afterwards to refresh the prompt files and configurations to the latest version.

3. Review a Pull Request

In Copilot / AI chat:

123 #review-pr

or use execute to force tool invocation (recommended for smaller models that may not call tools autonomously):

execute 123 #review-pr

Where 123 is the Azure DevOps or GitHub Pull Request number.

4. Self-review local changes

#self-review

Works offline — no Azure DevOps connection required.

⚙️ Gateway Token Limit

By default, REBUSS.Pure ships with GatewayMaxTokens set to 128 000 in appsettings.json. This hard cap matches the context window limit imposed by GitHub Copilot's proxy and prevents model_max_prompt_tokens_exceeded errors — even when the model's native context window is larger (e.g. Claude's 200K).

If you use Claude Code, Anthropic API directly, or any other client without a gateway limit, remove or disable this cap so you can use the model's full context window:

In appsettings.Local.json (next to the server executable):

{
  "ContextWindow": {
    "GatewayMaxTokens": null
  }
}

Or via environment variable:

ContextWindow__GatewayMaxTokens=0

When GatewayMaxTokens is null or 0, the limit is disabled and the full model-native context window from the ModelRegistry is used.