claudeplugins.
A complete site for Claude Code plugins — from setup to production-ready.
Articles — Page 2
20 postsClaude Code Agent Definitions: When YAML Beats Skills and Commands
A practical guide to Claude Code agent definition YAML — when to reach for an agent vs a skill vs a slash command, plus scoping conventions that keep your .claude directory sane.
7 minMCP Tool Registration: Multi-Tool Stdio Servers for Claude
Register multiple tools on a single MCP stdio server with strict input schemas, typed outputs, and a clean dispatch pattern Claude can call reliably.
6 minClaude Code Skill Allowed-Tools: Restrict Automation Surface
Lock down which tools a Claude Code skill can invoke using the allowed-tools YAML field. Patterns, gotchas, and a safer-automation checklist.
7 minValidating Claude Skill Inputs: A Contract-First Pattern
Build Claude skills that fail loudly on missing inputs and silently default safe ones. A contract-first validation pattern with required-vs-optional fields and null-safe defaults.
6 minThe Slash Command YAML Contract: Frontmatter Fields That Actually Matter
A field-by-field breakdown of name, description, and allowed-tools in Claude Code slash command frontmatter, with shipped examples.
6 minClaude Plugins Pattern: A Skill That Reads Files Before Writing Code
A production-tested Claude Code skill pattern that forces file reads before edits, with the YAML frontmatter, trigger language, and tool restrictions that make it reliable.
6 minBuilding an MCP stdio Server in Python with asyncio
Step-by-step tutorial for building a Model Context Protocol stdio server in Python using asyncio, with JSON-RPC framing, tool registration, and Claude Code integration.
5 minClaude Code Skill YAML: The Four-Field Frontmatter Contract
A field-by-field breakdown of the YAML frontmatter that turns a Markdown file into a Claude Code skill: name, description, allowed-tools, and model.
7 min