Anthropic's official "Code with Claude" workshop materials — nine hands-on courses on skills, MCP, evals and Claude Managed Agents.
- official
- workshops
- anthropic
- managed-agents
Claude Code Best Practices
Anthropic's guide to getting the most out of Claude Code.
Anthropic engineering's lessons on prompts, hooks, sub-agents, slash commands, and workflow design with Claude Code.
What it does
- Opens with the practice the page calls the difference between a session you watch and one you walk away from: give Claude a check it can run — tests, a build, a screenshot to compare
- Sequences the workflow: explore first, then plan, then code — plan mode separates exploration from implementation so the agent does not solve the wrong problem
- Covers the full configuration surface: specific prompts, rich content, environment setup, an effective CLAUDE.md (started with /init), and permissions
- Extension practices each get a section: CLI tools, MCP servers, hooks, skills, custom subagents, and plugins — plus communication patterns like asking codebase questions and letting Claude interview you
When to use it
The official best-practices page for Claude Code — send new users here first, and use it as a tuning checklist for an existing setup (permissions, hooks, CLAUDE.md hygiene).
Getting started
No install — a documentation page. The Claude Code docs expose a complete documentation index at /docs/llms.txt for discovering every related page.
Things to know
Practices evolve with Claude Code features (plan mode, hooks, subagents) — re-read after major releases.
Charted
Related on the map
Anthropic's announcement of Fable 5.1 and Mythos 5.1 — one base model, two guardrail postures.
- official
- models
- announcement
- fable
Anthropic's canonical essay on agent design — workflows vs agents, and five composable patterns.
- official
- agents
- engineering
- patterns