Mcp builder

Skill charlieviettq/awesome-agent-skill/.claude/skills/mcp-builder

Curated skill pack for LLM agents in engineer and science workflow (Cursor & Claude ready).

Install

npx -y skills add charlieviettq/awesome-agent-skill --skill mcp-builder

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

22 stars22 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Design and implement MCP servers—tools, resources, prompts, schemas, and error contracts for agent clients. Use when adding MCP integration or exposing capabilities to Cursor/Claude agents.

SKILL.md

1.7 KB, 337 tokens by cl100k_base, as published. Nobody here has run it

MCP builder

Design principles

One capability per tool — clear name, single purpose.
Strict schemas — JSON Schema for inputs; document required vs optional.
Predictable errors — machine-readable codes + human message; no stack traces to client by default.
Idempotent reads — writes clearly labeled; confirm destructive ops in description.

Server checklist

server identity and version documented
Tool list small and discoverable
Input validation before side effects
Timeouts and size limits on responses
Auth documented (env vars, OAuth, local only)
README: install, config, example tool calls

Tool definition pattern

{
  "name": "search_docs",
  "description": "Search internal docs by query. Read-only.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "query": { "type": "string", "description": "Search terms" },
      "limit": { "type": "integer", "minimum": 1, "maximum": 50 }
    },
    "required": ["query"]
  }
}

Testing

Unit test handlers with fixture inputs.
Integration test with MCP client or inspector.
Verify error paths (invalid args, upstream failure).

Security

No broad filesystem or shell unless required and scoped.
Sanitize paths; block path traversal.
Do not return secrets in tool results.

Use agent-tool-contracts for cross-platform tool design beyond MCP.