What it does This repository catalogs Claude Code’s sprawling feature set—subagents, slash commands, skills, MCP servers, hooks, memory rules, and a growing list of beta features like ultrareview and auto mode—and pairs each one with a best-practice guide and an implementation file you can actually run. Think of it as living documentation maintained by someone who uses the tool daily.

The interesting bit The repo is structured as a lookup table: each row links a feature to its official docs, a .md best-practice writeup, and a concrete implementation in the repo. That sounds obvious, but most tool documentation skips the middle step, leaving you to reverse-engineer “the right way” from marketing copy. Here, the middle step is the whole point.

Key highlights

• Covers 20+ features, from stable (commands, skills, settings) to beta (channels, ultraplan, agent teams)
• Each feature tagged with “Best Practice” and “Implemented” badges that link to actual files
• Includes orchestration workflows, e.g. a weather-orchestrator command showing multi-step agent coordination
• Tracks beta features with explicit beta tags so you know what might break
• Maintained with Claude Code itself—badge shows the exact version and timestamp of the last update

Caveats

• Heavy on badges, animated GIFs, and social-media embeds; the signal-to-noise ratio in the README is low
• Many “implementations” are thin markdown files rather than full working code
• Some features (checkpointing, devcontainers) are listed with no description at all

Verdict Worth bookmarking if you’re actively using Claude Code and tired of digging through official docs for the practical details. Skip it if you want deep architectural reasoning or reusable libraries—this is a curated index, not a framework.