What problem does it solve? Choosing, installing, and securely configuring Model Context Protocol servers for AI coding agents is confusing and risky, with tool bloat, credential exposure, and production-write dangers. This Skill provides a curated reference covering 50+ MCP servers with installation commands, security rules, and troubleshooting steps. ## Core Features & Use Cases - Category-Based Server Selection: Covers 9 categories including core development, databases, ops, productivity, fintech, web3, trading, creator tools, and reasoning/memory, with per-server install commands and environment variables. - Security Best Practices: Enforces three critical rules (limit to 5-7 servers, treat servers as untrusted code, never enable production writes) plus token scoping, read-only database users, and filesystem isolation patterns. - Troubleshooting & Configuration Examples: Diagnoses startup failures, authentication errors, and performance degradation, and provides ready-made JSON configs for dev, creator, and team productivity setups. - Use Case: A developer setting up Claude Code for the first time uses this Skill to pick a minimal 5-server dev setup (GitHub, filesystem, Postgres read-only, Brave Search, sequential thinking) with scoped tokens and pinned versions. ## Quick Start Ask the agent to show you the essential MCP servers for a minimal development setup and generate the configuration JSON for Claude Code.
