MCP Servers Focused on Security and Cryptography: A Complete Guide

The punkpeye/awesome-mcp-servers repository hosts multiple MCP servers specializing in cryptographic operations, identity verification, and post-quantum security analysis that AI agents can invoke via HTTP or stdio transports.

The Model Context Protocol (MCP) enables AI systems to discover and execute external tools through standardized interfaces. Within the punkpeye/awesome-mcp-servers repository, the cryptography category provides production-ready MCP servers that handle everything from Ed25519 signature verification to quantum-vulnerability scanning, allowing agents to perform security-critical operations without direct access to cryptographic libraries.

Available Security and Cryptography MCP Servers

The README.md file in the awesome-mcp-servers repository defines a dedicated "🔑 Cryptography" section that catalogs several specialized servers:

Ed25519 Identity Verification and Seal Verification

The Waxseal MCP server (npx @waxseal/mcp) provides on-chain Ed25519 identity management and seal verification. This tool generates and verifies Ed25519 signatures suitable for tamper-proof actions and human-signed approvals. It supports both Streamable HTTP and local stdio transports, making it accessible for web-based agents and local CLI workflows alike.

Enigma-Style Encryption

For educational and historical cryptography scenarios, the enigma-python-mcp server (npx enigma-python-mcp) emulates rotor-based Enigma machines. This server allows AI agents to encrypt and decrypt messages using the classic Enigma algorithm via HTTP transport, useful for demonstrating cryptographic concepts or handling legacy encryption formats.

Bitcoin-Anchored Verification Ledger (JIDEC)

The QuantaKrypto JIDEC server provides trust-less data integrity verification through Bitcoin anchoring. When invoked at https://mcp.quantakrypto.com, this server retrieves transaction raw bytes, recomputes SHA-256 hashes, and validates OpenTimestamp proofs on the Bitcoin blockchain. This creates cryptographically verifiable audit trails without requiring agents to run local Bitcoin infrastructure.

Byte-Wrangling Cryptographic Toolbox

mcp-bytesmith (uvx mcp-bytesmith) offers a comprehensive local cryptographic utility that requires no network calls. This stdio-based server provides:

  • Hex, Base64, and Base58 encoding/decoding
  • Cryptographic and CRC hashing algorithms
  • Ethereum-specific utilities (Keccak, EIP-191, EIP-712)
  • Cryptographically secure random token generation (CSPRNG)

Post-Quantum Readiness Scanning

The @quantakrypto/mcp server (npx @quantakrypto/mcp) scans source codebases for quantum-vulnerable cryptographic primitives such as RSA and ECDH. It suggests migrations to NIST-approved post-quantum algorithms including ML-KEM, ML-DSA, and SLH-DSA, and can verify implemented fixes. This server supports both HTTP and stdio transports for integration into CI/CD pipelines.

Architectural Patterns in Security MCP Servers

These MCP servers share a consistent architecture defined in the underlying repositories and referenced in CONTRIBUTING.md:

MCP Server Core

Each server implements a lightweight runtime using Node.js, Python, Deno, or compiled binaries. The core registers tool definitions using the MCP schema format, exposing cryptographic functions as discoverable tools that adhere to the Model Context Protocol specification.

Transport Layer Implementations

Security MCP servers support dual transport modes:

  1. Streamable HTTP: Compatible with MCP "list/get/call" RPC over HTTP endpoints
  2. stdio: CLI-based interfaces that read JSON-RPC messages from STDIN and write responses to STDOUT

This dual-mode approach allows agents to use local execution for sensitive operations (preventing data leakage) or remote calls for distributed verification.

Security-First Design Principles

According to the repository's CONTRIBUTING.md guidelines and individual server implementations, these tools enforce:

  • Read-only operations where possible to prevent accidental data modification
  • Strict input validation including AST parsing and strict parameter typing for cryptographic functions
  • Tamper-evident outputs via optional Ed25519 or on-chain signatures for verification results

Practical Code Examples

The following examples demonstrate how to invoke these security-focused MCP servers using standard MCP call patterns.

Verifying On-Chain Ed25519 Seals

To verify a cryptographic seal using the Waxseal server via HTTP transport:

curl -X POST https://api.waxseal.id/mcp/verifySeal \
  -H "Content-Type: application/json" \
  -d '{
        "seal": "0x01…",
        "message": "Transfer $1000"
      }'

This returns a structured JSON response indicating signature validity, which agents can parse directly into downstream workflows.

Generating Cryptographic Hashes Locally

For offline hash generation using the byte-wrangling toolbox:

uvx mcp-bytesmith hash \
  --algorithm sha256 \
  --data "Important payload"

The server returns JSON output such as {"hash":"a3f5…"}, allowing programmatic consumption without network latency or security risks.

Scanning for Quantum Vulnerabilities

To audit a codebase for post-quantum cryptographic risks:

npx @quantakrypto/mcp scan \
  --repo /path/to/project \
  --output json

This produces structured output like {"issues":[{"file":"src/crypto.py","type":"RSA","severity":"high"}]}, enabling automated security reporting in AI agent workflows.

Key Source Files and Repositories

The ecosystem of security MCP servers is organized across several key locations:

  • README.md (https://github.com/punkpeye/awesome-mcp-servers/blob/main/README.md): Contains the central "🔑 Cryptography" catalog with links, transport details, and descriptions for each security-focused MCP server
  • CONTRIBUTING.md (https://github.com/punkpeye/awesome-mcp-servers/blob/main/CONTRIBUTING.md): Defines security requirements for new server submissions, including input validation and audit logging standards
  • Individual Server Repositories:
    • waxseal-sdk: Implementation of Ed25519 verification tools
    • enigma-python-mcp: Enigma machine emulation server
    • quantakrypto/pqc-tools: Post-quantum cryptographic analysis tools

These resources collectively provide a complete ecosystem for integrating cryptographic security into AI agent workflows through standardized MCP interfaces.

Summary

  • MCP servers focused on security and cryptography provide AI agents with standardized access to cryptographic primitives without requiring direct library integration
  • The punkpeye/awesome-mcp-servers repository catalogs five major categories: Ed25519 identity verification, Enigma encryption, Bitcoin-anchored verification, byte-wrangling utilities, and post-quantum scanning
  • These servers support both HTTP and stdio transports, enabling both cloud-based and air-gapped secure operations
  • All implementations follow MCP schema standards and return structured JSON for automated processing
  • Security-focused design patterns include read-only defaults, strict input validation, and optional tamper-evident signatures

Frequently Asked Questions

How do I choose between HTTP and stdio transport for security MCP servers?

Use stdio transport for sensitive operations involving private keys or proprietary data, as it runs locally without network exposure. HTTP transport works better for distributed verification services like Bitcoin timestamp checking or when integrating with containerized environments. Both transports use identical MCP JSON-RPC schemas, so switching requires only configuration changes, not code modifications.

Can these MCP servers handle production-grade cryptographic operations?

Yes, but with important caveats. Servers like mcp-bytesmith and @waxseal/mcp implement proper cryptographic primitives suitable for production use. However, always verify that the specific server version adheres to current security standards (e.g., NIST recommendations) and runs in a trusted execution environment. The CONTRIBUTING.md file in the awesome-mcp-servers repository outlines security requirements for listed servers, but conduct your own audit before processing highly sensitive data.

What is the difference between the Waxseal and QuantaKrypto servers?

Waxseal specializes in Ed25519 identity verification and seal management, focusing on cryptographic signatures for approval workflows. QuantaKrypto provides Bitcoin-anchored verification and post-quantum analysis, handling blockchain-based integrity proofs and vulnerability scanning. While both handle cryptographic functions, Waxseal emphasizes identity and signing, whereas QuantaKrypto emphasizes audit trails and future-proofing against quantum attacks.

Do I need to install these servers globally to use them with AI agents?

No, most security MCP servers support ephemeral execution. Tools like uvx mcp-bytesmith and npx @quantakrypto/mcp run without permanent installation using package executors (uvx for Python, npx for Node.js). This approach reduces attack surface by ensuring you always run the latest version without maintaining local installations, though you can install them permanently if your environment requires offline operation.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"

Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →