MCP servers for an AI coding agent
The MCP server bundle to build an AI coding assistant that can read your repo and write code.
3 servers · ~22 tools · Last updated June 17, 2026
TL;DR: A capable coding agent needs to see your repository, touch local files, and reference accurate, version-correct documentation. This bundle gives it all three — the difference between an agent that guesses APIs and one that gets them right.
Bottom line: start with GitHub MCP Server and add the rest as your needs grow. All 3 install together via the merged config below (~22 tools total).
Tool budget: this stack exposes about 22 tools. That's within Cursor's practical ~40-tool ceiling, so all servers can stay enabled together. Check your own config →
What's in the stack
Read/write local project files (scope it to the repo directory).
Source ↗One-click config
All 3 servers merged into a single block — pick your client and paste.
{
"mcpServers": {
"github-mcp-server": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"GITHUB_PERSONAL_ACCESS_TOKEN",
"ghcr.io/github/github-mcp-server"
],
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "<your-github-pat>"
}
},
"filesystem-reference": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-filesystem",
"/absolute/path/to/dir"
]
},
"context7-mcp-server": {
"command": "npx",
"args": [
"-y",
"@upstash/context7-mcp"
],
"env": {
"CONTEXT7_API_KEY": "<optional-your-context7-api-key>"
}
}
}
}Paste into ~/Library/Application Support/Claude/claude_desktop_config.json and fully restart Claude Desktop. Replace any placeholder keys/tokens with your own.
Capabilities this stack covers
FAQ
Why add a docs server to a coding agent?
Training data lags releases. A docs MCP injects up-to-date API docs so the agent stops hallucinating methods that don't exist.
Is the filesystem server safe?
Scope it to your project directory, never your home folder. That keeps the agent's file access bounded.