# Itamos MCP Tools [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/Itamos-technologia/itamos-mcp-tools  
**GitHub Stars:** 1  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/itamos-mcp-tools

## Description
Code-map tools for AI agents: see a repo's structure first, then edit only what matters.

## Tools
Capabilities this server exposes over MCP:

- **read_file** — The single entry point. With no path it lists your projects; a project name shows its structure with coordinates (no code); a coordinate opens a file map or one segment in a single call. Edits one segment at a time, and every save is verified automatically.
- **master_architect** — The engine behind read_file. Indexes the codebase into a graph (files, imports, databases, servers) and keeps it in step with every save. In the sandbox it is not called directly: read_file does it.
- **write_file** — Creates new files only and never overwrites; edits go through read_file. Every new file belongs to a project the agent names: an existing one, or a new name to start one. Nothing is placed automatically.
- **git** — The original Itamos git tool, sandboxed for security: clone (remote URLs only, depth=1), status, commit. A clone is scanned into a project as soon as it finishes. Everything stays inside the session workspace; push is not available.
- **web_skeleton** — LLM-first web perception. 97% token reduction vs raw HTML. Actions: search, skeleton, read, click.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "itamos-mcp-tools": {
    "url": "https://mcp.itamos-technologia.com/mcp"
  }
}
```

## Documentation

## What Itamos MCP Tools MCP server does

Itamos MCP Tools MCP server provides MCP tools for navigating and modifying codebases without returning an entire repository as raw text. It also supports shallow repository cloning, Git status and commits, and structured web-page access. Each client receives an isolated workspace in the hosted sandbox, and inactive workspaces are removed after 10 minutes.

The central workflow moves from project discovery to a file map and then to a selected segment. A project view shows folders and files with coordinates but not source code. A file map lists named segments, such as classes and methods, so an agent can request only the section relevant to a task.

## How it works

The `read_file` tool is the primary interface. With no path, it lists projects. Supplying a project name returns its structure, while adding a coordinate opens a file map or a specific segment. Searches can operate on project paths, file names, folders, or segment names. A trace option exposes file connections or project data flow.

`master_architect` builds a graph of files, imports, databases, and servers and updates that graph after saves. In the sandbox, agents use it indirectly through `read_file`. Files are parsed into named segments, and large structures return an overview with a request for a search term rather than an unbounded dump.

Edits are performed one segment at a time and only after the agent has read that segment since the previous change. Save checks cover syntax, selected configuration validators, linters, local imports, and SQL against the live database schema where supported. Successful saves are re-indexed.

## Setup and configuration

The hosted endpoint is `https://mcp.itamos-technologia.com/mcp`, which can be added as a remote MCP server. The README also documents Claude.ai connector use and direct API integration. For self-hosting, use Linux with ZFS, Node.js 20 or newer, native-module build tools, Git, and Chrome or Chromium. Install dependencies with `npm install`, create sandbox slots with the provided script, and start the service with `SANDBOX_ROOT` and `SANDBOX_TOTAL_SLOTS` set.

A local deployment listens at `http://localhost:4200/mcp` by default. Public deployments require firewall and reverse-proxy configuration; direct access to port 4200 should not be exposed externally.

## Tools and capabilities

- `read_file`: lists projects, maps structures and segments, reads targeted code, searches, traces connections, and applies verified segment edits.
- `master_architect`: maintains the codebase graph used by navigation.
- `write_file`: creates new files without overwriting existing files.
- `git`: performs remote-URL shallow clones, status checks, and commits inside the session workspace.
- `web_skeleton`: searches, summarizes page structure, reads selected sections, and follows links with `search`, `skeleton`, `read`, and `click` actions.

## Limitations and notes

Itamos MCP Tools MCP server does not provide Git push. Clones use remote URLs and depth one. New files are assigned to a project named by the agent, and nothing is placed automatically. Hosted workspaces are temporary and should not be used for sensitive data. Self-hosting requires ZFS; the service refuses to start without it. Optional summarization works with an OpenAI-compatible chat endpoint, but the core tools still work without one.

_Full upstream README: https://allmcps.com/mcp/itamos-mcp-tools/readme_