Skip to main content
MCP server CX Codex

mcp-server · wasmagent — Claude Code MCP server: what it does, how to install

Moves tool calls into a WASM sandbox powered by QuickJS. It reduces context token usage by presenting just docs_search and execute_code to the model.

Install into Claude Code · Platform CLI All methods and platforms ↓
claude mcp add mcp-server -- npx -y @wasmagent/mcp-server
Open on GitHub

What mcp-server does

The server packages tool collections behind a minimal two-method surface. Instead of serializing dozens of full schemas into the prompt, the host model sees only search and execution endpoints. The model then writes a small script to dispatch whichever downstream methods it needs. Under 30 downstream utilities, initial prompt token overhead drops to 13.6%.

Execution takes place inside an isolated WebAssembly QuickJS kernel. A CapabilityManifest controls access boundaries: cpuMs limits, memoryLimitBytes caps, plus strict allowedHosts and allowedPaths entries. Network requests to unapproved destinations fail immediately.

This project is in beta. Launching the zero-config binary via npx provides an empty inspection server with no tools attached. Practical deployments require a custom Node entrypoint configuring ToolRegistry.

What it helps with

  • shrink prompt token overhead for large tool registries
  • isolate model-generated code in a WebAssembly sandbox
  • enforce strict network and disk permissions with a manifest
  • host an MCP agent across stdio or edge HTTP runtimes

Example prompts for the agent

  • Execute the data transformation inside execute_code with no network access
  • Look up registered utilities using docs_search before running the script
  • Run this calculation under a 5000ms CPU execution cap

How to install mcp-server

Platform CLI

claude mcp add mcp-server -- npx -y @wasmagent/mcp-server

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about mcp-server

How to install mcp-server in Claude Code?

Run claude mcp add mcp-server -- npx -y @wasmagent/mcp-server. This command attaches the server using the stdio transport.

What is the purpose of code-mode?

It replaces heavy tool schemas with just docs_search and execute_code. The assistant inspects tool docs and runs code locally, preserving context budget.

How does the server enforce security boundaries?

The runtime checks CapabilityManifest inside the QuickJS environment. Any script attempting network calls outside allowedHosts triggers a runtime exception.

Similar MCP servers

All MCP servers for Agent setup: memory, context, orchestration →

mcp-server by other authors comes first.

mcp-server HeyPuter Connects your AI agent to Puter for files, websites and workers. Useful when the project already lives in Puter's cloud. CX Codex +1 ★ 43,902repo stars mcp-server mapbox Connects Mapbox location APIs to your AI assistant for geocoding, multi-modal routing, and map rendering. Helpful when building travel planners or spatial query assistants. CX Codex +1 ★ 363repo stars mcp-server PackmindHub Connects team engineering playbooks and shared coding standards from Packmind directly to AI coding agents. CX Codex +1 ★ 319repo stars mcp-server BingoWon Search Apple developer documentation and WWDC transcripts, with results ranked for relevance to your question. Useful when an agent needs API references or coding context. CX Codex +1 ★ 121repo stars mcp-server UI5 Connects SAPUI5 and OpenUI5 developer tooling to AI assistants. Helps scaffold apps, inspect manifest schemas, and run linters directly in chat. CX Codex +1 ★ 105repo stars REA morluto Analyzes binaries, Electron bundles, and JavaScript runtimes directly from your agent. It inspects compiled apps without source code to recreate specific features. CX Codex +1 ★ 392,262repo stars GitHub official github Wires your AI agent into GitHub: read code, open issues and pull requests, watch Actions runs. Saves a trip to the browser. CX Codex +1 ★ 33,989repo stars Netdata netdata Hands your AI agent live monitoring data: metrics, logs, alerts and anomaly detection. Useful when you need to know what a server is doing without opening the dashboard. CX Codex +1 ★ 81,163repo stars browser-use browser-use Give an agent browser tasks such as finding information, extracting page data, and filling forms. Connect it when those actions need to happen on the web. CX Codex +1 ★ 119,231repo stars Codebase Memory DeusData A local MCP server that maps your repo's code, docs and infrastructure into a graph. Gives the agent structural context without re-reading files. CX Codex +1 ★ 47,577repo stars

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor