Configure and build Model Context Protocol (MCP) servers for Claude Code integration. Set up database, filesystem, git, and API connections. Build custom MCP servers with TypeScript/Python SDK, implement tools and resources, configure transports (stdio, HTTP), and deploy for production.
Install
npx skillscat add vasilyu1983/ai-agents-public/claude-code-mcp Install via the SkillsCat registry.
This skill enables configuration and deployment of Model Context Protocol (MCP) servers for integrating Claude Code with databases, filesystems, Git repositories, and APIs. It addresses the need to securely connect AI agents to structured data sources and remote services via standardized interfaces. Use it when building custom MCP servers in TypeScript/Python or deploying prebuilt servers for production workflows requiring data access and tool orchestration.
Claude Code MCP
Specification: https://modelcontextprotocol.io/specification/2025-11-25 (November 2025)
Use this skill when you need to:
- Configure
.claude/.mcp.jsonfor official/community MCP servers - Build custom MCP servers (TypeScript or Python)
- Deploy and harden remote MCP servers (HTTP transport)
Quick Start (Local stdio via npx)
- Create or edit
.claude/.mcp.json:
{
"mcpServers": {
"postgres": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-postgres"],
"env": { "POSTGRES_URL": "${DATABASE_URL}" }
}
}
}- Provide env vars and validate the connection:
export DATABASE_URL="postgresql://user:pass@localhost:5432/db"
claude mcp list
claude mcp get postgresNotes:
- Default config path is typically
.claude/.mcp.json. If needed, override viaclaude --mcp-config <path>. - Prefer
python3as the Python interpreter in server configs unless your environment guaranteespython.
Permission Management (Recommended)
# Allow all tools from a server (wildcard)
claude mcp add --allow "mcp__postgres__*" postgres -- npx -y @modelcontextprotocol/server-postgres
# Allow specific tools only
claude mcp add --allow "mcp__postgres__query,mcp__postgres__list_tables" postgres -- npx -y @modelcontextprotocol/server-postgres
# Deny a specific tool
claude mcp add --deny "mcp__filesystem__write_file" filesystem -- npx -y @modelcontextprotocol/server-filesystem ./dataProduction Guardrails (Required)
- Assume tool outputs are untrusted (prompt injection). Sanitize/structure before reuse.
- Default to least privilege: read-only DB, scoped filesystem allowlists, minimal tool allowlists.
- Keep secrets out of
.mcp.json; inject via env vars or a secret manager at runtime. - Add timeouts, retries, and rate limits; log all tool invocations for audit.
What To Read Next
- Choose an existing server:
references/mcp-servers.md - Build a custom server:
references/mcp-custom.md - Implementation patterns (DB/API/filesystem):
references/mcp-patterns.md - Security hardening (OAuth, scopes, injection defense):
references/mcp-security.md - Templates:
assets/database/template-mcp-database.md,assets/filesystem/template-mcp-filesystem.md,assets/api/template-mcp-api.md,assets/deployment/template-mcp-docker.md - Curated links:
data/sources.json