Claude Code Routines Mastery
Master Claude Code Routines in 2 hours. Build scheduled, GitHub-event, and API-triggered routines that run on Anthropic's cloud while your laptop's closed. 8 lessons, certificate included.
Claude Code Routines launched on April 14, 2026 — and within 24 hours, half the community was hitting quota walls by noon. The other half had figured out the discipline that separates routines that scale from routines that burn out overnight.
This course teaches you that discipline, in two hours, with three real routines shipped to your own account by the time you finish.
What you’ll actually build:
- A scheduled backlog triage routine that runs weekdays at 9am, reading Linear and producing a 300-word morning report — using only the Linear MCP connector, nothing else loaded.
- A PR review routine that fires on GitHub pull_request events, filtered to your main branch and
needs-reviewlabel, maintaining one session per PR so follow-up commits build on prior feedback. - A Stripe webhook → onboarding draft email routine, wired through a secure proxy, with least-privilege Gmail scopes that can only create drafts (never send).
What makes this course different from the docs:
The Anthropic launch docs explain what Routines does. This course explains what everyone learned in the first week about how to use it without blowing your quota, leaking scope, or shipping something that silently fails at day 7.
Each lesson ends with the anti-patterns that tripped up real users in the first 24 hours — the unfiltered triggers, the chatty prompts that burned 56% of their tokens on narration, the “all connectors enabled by default” trap, the OAuth tokens that expired silently after two weeks. You’ll learn the fixes alongside the patterns.
Who this course is for:
Engineers who’ve used Claude Code at least a few times and want to move from “AI pair programming” to “AI cloud of agents running real work.” You don’t need deep DevOps experience — just a working Claude Code install, a Pro or Max plan, and one real workflow you’d like to automate.
By the end, you’ll be shipping routines the same week Routines launched. That’s a narrow window, and it’s now.
What You'll Learn
- Explain how Routines differ from interactive Claude Code, Cowork, and /loop
- Build scheduled, GitHub-event, and API-triggered routines with proper filter discipline
- Configure MCP connectors (Notion, Slack, Gmail, Linear, Postgres) with least-privilege OAuth scopes
- Apply cost-engineering patterns to stay inside Pro 5/day, Max 15/day, or Team 25/day caps
- Evaluate when Routines is the right tool vs GitHub Actions, Zapier, n8n, or cron
- Ship three production-quality routines to your own account with a dry-run and iteration workflow
After This Course, You Can
What You'll Build
Course Syllabus
Prerequisites
- Basic familiarity with Claude Code (you've run at least a few interactive sessions)
- Pro, Max, Team, or Enterprise Claude plan (Routines aren't available on Free)
- Comfort with terminal, git, and basic JSON
Frequently Asked Questions
What's the difference between Routines and interactive Claude Code?
Interactive Claude Code is pair-programming — you direct Claude in real time. Routines run unattended on Anthropic's cloud, triggered by a schedule, a GitHub event, or an API call. Your laptop can be closed.
Do I need a paid Claude plan?
Yes. Routines are available on Pro, Max, Team, and Enterprise plans. Daily run caps vary: Pro gets 5/day, Max 15/day, Team and Enterprise 25/day per user.
Is this course current with the April 2026 launch?
Yes. The course was built the week Routines launched (April 14, 2026) based on Anthropic's launch docs and first-day community experience. We flag anything that's still in flux in the research preview.
Will the routines I build work forever?
The core patterns you learn (filter discipline, prompt structure, connector scoping, cost engineering) are timeless. Specific UI details may shift as Anthropic iterates — we focus on the patterns that survive platform changes.
