# @sebastienrousseau/crypto-mcp [Health: Active]

**Category:** 🔑 Cryptography  
**Repository:** https://github.com/sebastienrousseau/crypto-service  
**GitHub Stars:** 9  
**npm Downloads (last month):** 2633  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/sebastienrousseau-crypto-mcp

## Description
Model Context Protocol (MCP) server that gives AI assistants and agents cryptography tools: key generation, encryption, ML-KEM, ML-DSA signing, hashing, streaming encryption, KMS and CBOM.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What @sebastienrousseau/crypto-mcp does

@sebastienrousseau/crypto-mcp makes cryptographic functions available as MCP tools for AI assistants and agents. Its surface covers common operations such as hashing, encryption, and key generation, alongside post-quantum mechanisms including ML-KEM key encapsulation and ML-DSA signing.

The server is part of the Crypto Service ecosystem, a TypeScript cryptography project that also provides library, server, command-line, and related packages. The broader toolkit lists more than 50 algorithms and is intended for use across Node.js, browser, and Edge environments. This listing focuses on the MCP interface for agent-driven access to those capabilities.

## Tools and capabilities

Supported areas include:

- Generating cryptographic keys.
- Encrypting and decrypting data.
- Encrypting data as a stream.
- Calculating hashes, including algorithms such as SHA-256 and BLAKE3.
- Using ML-KEM for post-quantum key encapsulation.
- Creating and verifying ML-DSA signatures.
- Working with key-management-system operations.
- Producing or handling cryptographic bill of materials (CBOM) information.

The underlying toolkit also demonstrates authenticated encryption with AES-256-GCM and identifies XChaCha20-Poly1305 among its supported usage examples. Post-quantum examples include ML-KEM parameter set 768 and ML-DSA signing. These capabilities make the server relevant to agents handling integrity checks, protected payloads, signatures, key exchange, and cryptographic inventory tasks.

## How it works

An MCP-capable assistant selects the relevant cryptography operation and supplies its inputs through the server interface. The operation returns the resulting cryptographic data or status to the agent, allowing the assistant to continue a workflow involving encryption, hashing, key material, signatures, or inventory information.

The project is implemented in TypeScript and uses Node.js. Its post-quantum implementations come from `@noble/post-quantum`. The associated toolkit also offers direct library APIs and a command-line interface, but those are separate access paths from the MCP server.

## Setup and configuration

The project requires Node.js 22.0.0 or newer. Development uses pnpm 11.9.0, and the repository specifies TypeScript around version 5.9.3. The wider ecosystem can be installed from npm as scoped packages, built from the Git repository, or run from a published Docker image, but the MCP-specific installation command and launch configuration should be taken from the package and client setup used in a deployment.

## Limitations and notes

The post-quantum algorithms rely on `@noble/post-quantum`, which has not been independently audited and does not guarantee constant-time execution. The suite is not FIPS 140-3 validated. These considerations matter when deciding whether the server is suitable for production environments with formal validation, side-channel, or compliance requirements.

Cryptographic operations should be evaluated according to the sensitivity of the data, key-handling model, and trust placed in the MCP client and agent. The server provides access to cryptographic functions; it does not by itself establish a complete key-management or compliance architecture.

_Full upstream README: https://allmcps.com/mcp/sebastienrousseau-crypto-mcp/readme_