LABSCO SUMMARY

A whole codebase becomes one document an agent can read.

Point it at a directory or a GitHub URL and one packed file comes back with the tree, the contents and the metrics, instead of you opening files one at a time. It answers with an id rather than the pack, so you choose what part of it ever reaches the model.

What it is

The MCP side of Repomix, the packer that flattens a codebase into one file an AI can read. It runs locally over stdio, and it will pack a directory on your machine or clone a public repository and pack that. Packing runs a secret scan first, and the output is addressed by an id rather than returned whole.

What you get

A local directory packed into one consolidated file carrying the file tree, the contents and the metrics
A GitHub repository packed the same way — it fetches and clones it for you
A packed output read back by line range, or grepped with JavaScript regular expressions, returning matching lines with context
A packed output you already have attached, so a pack made elsewhere can be worked with here
A Claude Agent Skill generated from a directory or a remote repository — SKILL.md plus a references/ folder holding a summary, the project structure with line counts and more
A --sandbox mode that confines the file tools to one workspace root: absolute paths, ~, .., Windows drive and UNC paths are all refused, as are symlinks that resolve outside it
Two raw file tools that exist only in sandbox mode, where the workspace root is what bounds them — file_system_read_file, file_system_read_directory

Requirements

npx on your PATH — repomix --mcp starts MCP mode from the same package. Packing a remote repository clones it first, so that call needs network access. By default the server can read any path the host user can; --sandbox narrows that to one directory, and in exchange remote packing, skill generation and attaching external outputs are switched off. That confinement is at the tool layer, not the operating system's.

Setup effort

One command — repomix --mcp

See installation options

USE CASES

01

Reading a repository you did not write

Pack the GitHub URL and ask questions against the result. One document with the tree and the contents beats forty tabs, and the metrics tell you where the weight actually is.

pack_remote_repositoryread_repomix_output

02

Asking questions that span your codebase

Pack the local directory, then grep the pack with a regular expression and read the matching lines with their context. That is how you find every call site of something without a language server.

pack_codebasegrep_repomix_outputread_repomix_output

03

Fitting a large repository into a session

The pack comes back as an id, not as content, so you pull line ranges and grep hits rather than pushing the whole thing into context. Compression and per-file inclusion levels narrow it further before it is ever read.

pack_codebaseread_repomix_outputgrep_repomix_output

04

Handing an agent a repository without handing it your disk

Run with --sandbox pointed at one project. Every path is then resolved inside that root, and paths that try to climb out — through .., an absolute path or a symlink — are refused rather than followed.

pack_codebase

05

Turning a project into a reusable reference

Generate a Claude Agent Skill from the directory or a remote repository and get a SKILL.md plus a references folder with the summary and the project structure. It becomes something an agent loads on demand rather than a pack you paste each time.

06

Picking up a pack someone else made

Attach an existing packed output and read or grep it here. Useful when the packing happened in CI, or on a machine that could reach a repository yours cannot.

attach_packed_outputread_repomix_outputgrep_repomix_output

6 tools

Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.

`attach_packed_output`

Attach an existing Repomix packed output file for AI analysis. This tool accepts either a directory containing a repomix output file or a direct path to a packed repository file. Supports multiple formats: XML (structured with <file> tags), Markdown (human-readable with ## headers and code blocks), JSON (machine-readable with files as key-value pairs), and Plain text (simple format with separators). Calling the tool again with the same file path will refresh the content if the file has been updated. It will return in that case a new output ID and the updated content.

BEFORE YOU INSTALL

What to know first

Tools you get

6 tools when we connected.

INSTALLATION

Connect your client

Server setup

TERMINAL

repomix --mcp

MCP Clients

VS Code

VS CODE · TERMINAL

code --add-mcp '{"name":"repomix","command":"npx","args":["-y","repomix","--mcp"]}'

VS Code Insiders

VS CODE INSIDERS · TERMINAL

code-insiders --add-mcp '{"name":"repomix","command":"npx","args":["-y","repomix","--mcp"]}'

Cline

CLINE · JSON

{
  "mcpServers": {
    "repomix": {
      "command": "npx",
      "args": [
        "-y",
        "repomix",
        "--mcp"
      ]
    }
  }
}

Cursor

In Cursor, add a new MCP server from Cursor Settings > MCP > + Add new global MCP server with a configuration similar to Cline.

Claude Desktop

Edit the claude_desktop_config.json file with similar configuration to Cline's config.

Claude Code

CLAUDE CODE · TERMINAL

claude mcp add repomix -- npx -y repomix --mcp

Docker

DOCKER · JSON

{
  "mcpServers": {
    "repomix-docker": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/yamadashy/repomix",
        "--mcp"
      ]
    }
  }
}

Repomix

A whole codebase becomes one document an agent can read.

What it is

What you get

Requirements

Setup effort

Reading a repository you did not write

Asking questions that span your codebase

Fitting a large repository into a session

Handing an agent a repository without handing it your disk

Turning a project into a reusable reference

Picking up a pack someone else made

6 tools

`attach_packed_output`

`generate_skill`

`grep_repomix_output`

`pack_codebase`

`pack_remote_repository`

`read_repomix_output`

What to know first

Tools you get

Connect your client

Server setup

MCP Clients

VS Code

VS Code Insiders

Cline

Cursor

Claude Desktop

Claude Code

Docker

Repomix

A whole codebase becomes one document an agent can read.

What it is

What you get

Requirements

Setup effort

Reading a repository you did not write

Asking questions that span your codebase

Fitting a large repository into a session

Handing an agent a repository without handing it your disk

Turning a project into a reusable reference

Picking up a pack someone else made

6 tools

attach_packed_output

generate_skill

grep_repomix_output

pack_codebase

pack_remote_repository

read_repomix_output

What to know first

Tools you get

Connect your client

Server setup

MCP Clients

VS Code

VS Code Insiders

Cline

Cursor

Claude Desktop

Claude Code

Docker

Same category

Madar

Contextful

vnsh Ephemeral File Sharing

RepoRecall

`attach_packed_output`

`generate_skill`

`grep_repomix_output`

`pack_codebase`

`pack_remote_repository`

`read_repomix_output`