phab-arc-mcp
MCP server for reviewing Phabricator revisions. It fetches revision context and creates draft inline comments via Conduit API.
README
phab-arc-mcp
phab-arc-mcp is a small MCP stdio server for reviewing Phabricator revisions.
What This Server Exposes
The public MCP surface is intentionally small:
review-phabinline-comments-phab
What review-phab Does
review-phab is the tool you call when you want an AI reviewer to review a Differential properly.
You give it a revision ID like D35297. It fetches the revision title, summary, changed files, raw diff, and the linked Maniphest task context, then packages all of that together with the review prompt the model should use.
From a user point of view, this means you do not have to manually open the revision, copy the diff, inspect related tasks, and explain the background before asking for a review. The tool gives the model the code changes and the surrounding context in one shot.
That makes reviews more useful because the model can judge the change against the actual task being solved, not just the raw code diff. It can use the overall repository context to reason about likely bugs, regressions, and edge cases, and it can also check whether the revision actually does what the linked task says it should do. In practice.
After that, inline-comments-phab can take the review findings and create draft inline comments directly on the Differential.
Important details:
- comments are created as draft inlines, not published comments
Installation
Requirements
- Node.js 18 or newer
npm- access to a Phabricator instance with the Conduit API enabled
- a Conduit API token for an account that can read the revisions and tasks you want to review
1. Clone the repository
git clone https://github.com/wxomi/node-chat.git
cd mcp-phab
2. Install dependencies
npm install
3. Build the server
npm run build
This produces the MCP server entrypoint at dist/server.js.
4. Configure environment variables
Set these before starting the server or before wiring it into your MCP client:
PHAB_API_TOKEN- required
- primary token variable used by the server
CONDUIT_TOKEN- optional alias for
PHAB_API_TOKEN
- optional alias for
PHAB_CONDUIT_TOKEN- optional alias for
PHAB_API_TOKEN
- optional alias for
PHAB_CONDUIT_URI- optional
- default:
https://phab.instahyre.com/ - should point to the base URL of your Phabricator instance
PHAB_ARC_TIMEOUT_MS- optional
- default:
30000 - request timeout in milliseconds
Example:
export PHAB_CONDUIT_URI="https://phabricator.example.com/"
export PHAB_API_TOKEN="api-xxxxxxxxxxxxxxxx"
export PHAB_ARC_TIMEOUT_MS="30000"
If you do not set PHAB_CONDUIT_URI, the server will try https://phab.instahyre.com/.
5. Start the server manually (optional)
You usually do not need this step when using the server through an MCP client. In the normal setup, the MCP client launches the stdio server process for you using its mcpServers configuration.
Run the built server manually only if you want to verify that the binary starts correctly:
npm start
For local development without building on every change:
npm run dev
The server uses stdio transport, so manual npm start is optional and is not a required installation step.
Setup In An MCP Client
Codex / MCP config using the built server
{
"mcpServers": {
"phab-arc-mcp": {
"command": "node",
"args": ["/absolute/path/to/phab-arc-mcp/dist/server.js"],
"env": {
"PHAB_CONDUIT_URI": "https://phab.instahyre.com/",
"PHAB_API_TOKEN": "api-xxxxxxxxxxxxxxxx"
}
}
}
}
Setup Checklist
Use this to verify the installation quickly:
npm installcompletes successfullynpm run buildcreatesdist/server.js- your MCP client is configured with
PHAB_API_TOKEN PHAB_CONDUIT_URIpoints to the correct Phabricator base URL- the MCP client can see the
review-phabandinline-comments-phabtools
Example Usage
Step 1: get review context
Call review-phab with:
{
"revision_id": "D35297"
}
Notes
- This server talks to Phabricator over the Conduit HTTP API.
- A valid API token is required for every tool call.
- Comments created by
inline-comments-phabare draft inlines and are not published automatically.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.