claude-code-best-practice
A reference collection mapping Claude Code's own concepts and newer features, such as subagents, commands, skills, hooks, MCP servers and plugins, to where each one lives in the configuration and where to learn more.
A core concepts table lists subagents, commands, skills, workflows, hooks, MCP server configuration, plugins, settings, the status line, memory files, checkpointing, sessions and context-window commands, each with its file location and linked guides. A second "hot" table tracks newer or less-documented capabilities, including an ultra-review command, dev containers, channels, an auto-permission mode, power-ups, an advisor tool, computer use, the Agent SDK, a Ralph-Wiggum-style loop plugin, Chrome integration, the web version of Claude Code, artifacts, Slack integration, managed code review, GitHub Actions and remote control.
Each row links out to the relevant official or community documentation rather than duplicating it in full.
HTML, MIT licensed.