Security Headers MCP server

dev-mcp-servers

Ten small MCP servers for everyday development and security checks. Each one is a separate npm package you add to Claude Code, Claude Desktop, Cursor or any other MCP client with one line, and each does one job well: look up vulnerabilities, grade security headers, decode a JWT, test a regex, explain a cron expression, validate JSON against a schema, lint an OpenAPI document or a Dockerfile, read statistics out of a git repository, or build an llms.txt.

All servers are TypeScript on the official @modelcontextprotocol/sdk, speak stdio only, validate every input with zod, bound their inputs and run time, send no telemetry, and ship with node:test suites that cover every tool.

Servers

Every package README lists its tools with inputs and outputs, the exact claude mcp add command, a .mcp.json snippet, and what the server touches on disk and on the network.

Quick start

With Claude Code:

claude mcp add jwt-tools -- npx -y @basitalisandhu/mcp-jwt-tools@0.2.0
claude mcp list

Add -s user to install for every project rather than the current one. For clients that read .mcp.json (Claude Code, Claude Desktop, Cursor), add an entry per server:

{
  "mcpServers": {
    "jwt-tools": { "command": "npx", "args": ["-y", "@basitalisandhu/mcp-jwt-tools@0.2.0"] },
    "osv-advisories": { "command": "npx", "args": ["-y", "@basitalisandhu/mcp-osv-advisories@0.2.0"] }
  }
}

Shortened. The full README is on GitHub.