MCPC
An MCP builder agent that picks servers from the official registry and composes them into your own agentic MCP server.
- 105 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Composed agents inherit the access of all connected MCP servers
- Code-execution modes need separate isolation setup
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/mcpc-tech-mcpcDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Composed agents inherit the access of all connected MCP servers
- Code-execution modes need separate isolation setup
This is third-party code. Review the repository files before installing.
What it does
MCPC builds agentic MCP servers from existing tools. You describe a task in text, reference servers from the official MCP registry, and MCPC turns them into a new agent tool that works in any MCP client. You can change descriptions and parameters of other people's tools, wrap and filter results for your scenario, expose each agent as its own tool and compose them into a team. The entry point for agents is `mcpc-builder-agent`: add an entry running `npx -y @mcpc-tech/builder mcpc-builder-agent` to the client and the AI helps find servers and generate configuration. For developers there is the `@mcpc/core` SDK and the `@mcpc-tech/cli` CLI, which can start from a remote configuration. Execution modes include agentic, AI SDK sampling, ACP and sandboxed code execution. Skills following the Agent Skills specification are supported.
Who it is for. Developers who build their own MCP servers on top of existing ones and want to narrow other people's tools to a specific scenario.
Good fit when
- Several MCP servers need merging into one agent with a clear task.
- Another tool has an overly generic description and parameters.
- Call results must be filtered or wrapped before reaching the model.
- You need to sketch an agent quickly: pick servers from the registry and generate config.
Not a fit when
- One ready MCP server without composition is enough.
- The project does not use Node.js or Deno.
- Agentic sampling and ACP modes are overkill for a simple proxy.
Example request
Use mcpc-builder-agent to find GitHub and Slack servers in the registry and compose an agent that prepares a change digest.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
This is primarily an SDK and a set of execution modes, and the `mcpc-builder-agent` helper only eases the start. Agents built on other servers inherit their access, and code-execution modes require separate sandbox setup. Node.js or Deno is needed and, for sampling modes, a connected model. Documentation is in English. The repository was last updated on 6 October 2026 and uses the MIT licence.
How to disable. Remove the `mcpc-builder-agent` entry from the MCP client configuration. Delete servers and configurations you generated separately.
FAQ
What is an agentic MCP server?
A server whose tool is itself an agent: it is composed from a task description and references to other MCP tools and works in any client.
How to start the builder helper?
Add an entry with `npx -y @mcpc-tech/builder mcpc-builder-agent` to the client, after which the AI helps find servers and generate configuration.
Which packages exist for developers?
The `@mcpc/core` SDK (also `@mcpc-tech/core` on npm) and the `@mcpc-tech/cli` CLI, which starts a server from a remote configuration.
Which execution modes exist?
`agentic`, `ai_sampling`, `ai_acp`, `code_execution` and `code_execution_sampling`, each with its own implementation.
Are skills supported?
Yes, per the Agent Skills specification: they can be deployed, shared via MCP and declare tool dependencies.
Discussion
No questions or reviews yet.
Related
MCP serversA universal MCP CLI client from Apify: persistent sessions, OAuth, JSON output for scripts and shell access to any MCP server for agents
An MCP server that installs other MCP servers from npm and PyPI at the agent's request and writes them into the config
An MCP server that runs sub-agents defined in Markdown through Codex, Claude Code, Cursor CLI and other CLIs from any MCP client
An MCP server that turns a request into a step plan with dependencies and keeps context across agent sessions
An MCP server that inserts a human feedback step into agent tasks: a web or desktop window for text, images and prompts
A mail server for agents sharing one repository: messages, threads, file reservations and a 16-screen TUI
README in short
The README splits capabilities into three scenarios: the ready mcpc.tech site, an interactive agent and direct coding, then shows an example fork of a coding agent.