Once you understand how Claude Code works under the hood, the next step is learning how to steer it. In this workshop, you will learn how to shape Claude Code output using persistent configuration files, progressive disclosure patterns, and reusable skills that travel across projects.
What You Will Learn
- AGENTS.md files — how persistent markdown files give Claude Code memory of your preferences and coding conventions
- Steering with CLAUDE.md — using configuration rules to enforce patterns like object parameters over positional parameters
- Progressive disclosure — structuring instructions so only relevant context loads for each task, keeping the context window lean
- Agent skills — discoverable, scoped instruction sets that extend Claude Code capabilities without polluting global context
- Writing skills with skills — using a meta-skill to generate reusable migration skills you can share across your organization
- Automatic memory — how Claude Code writes and maintains its own steering documentation over time