This course covers Claude Code, Anthropic's agentic coding tool, with a focus on working hands-off. You'll learn to steer long-running sessions, configure rules Claude can't skip, automate recurring work, and verify results you didn't watch happen. Along the way you'll scope work with plan mode, keep CLAUDE.md files lean and followable, enforce guardrails with hooks, and run Claude Code from cloud routines or your own scripts. It all builds toward packaging a setup you trust as a plugin your whole team can install.

By the end of this course, you'll be able to: - Steer long Claude Code sessions using plan mode, directed /compact summaries, and rewind checkpoints - Run autonomous and parallel work with goals, loops, and git worktrees - Write lean, specific CLAUDE.md rules that Claude reliably follows - Build a verification skill that runs tests, reads diffs, and reports evidence automatically - Match Claude Code's six permission modes to the level of autonomy each task needs - Write hooks that enforce hard rules by allowing, denying, or rewriting Claude's actions - Automate recurring work with cloud routines, headless mode, and the GitHub Action - Review pull requests automatically with Code Review and apply findings locally - Verify unsupervised runs before shipping, and package your setup as a plugin for your team Prerequisites - Basic familiarity with command line interfaces and git - Access to Claude Code - Prior hands-on experience with Claude Code basics is helpful — this course focuses on working autonomously rather than first steps