Claude Code, beyond one developer
The Advanced course. Run sessions in parallel in worktrees, choose the orchestration pattern a job actually needs, turn Claude Code into a script and a CI step, share the setup as committed configuration — and put bounds on every run that works alone. Seven lessons and a capstone. About 37 minutes of reading, plus agent run time; every lesson ends with an exercise in your own repo.
The level at a glance
Beginner made one change safe, in one session, by one developer. This level is everything after that: several sessions at once, work nobody watches, and a setup a whole team runs — each step carrying the same review discipline, because unattended work is exactly where the discipline earns its keep. We ship this site through agent sessions held to these patterns; the habits below are the ones that survived.
You are ready if you can write a project skill, enforce a rule with a hook, delegate to a custom subagent, and connect an MCP server — the Intermediate toolkit, which installment 2 teaches end to end. If you skipped it, the Beginner exit is the honest floor for lessons 1, 4 and 5.
The exercises assume a repository you can experiment in; the GitHub Actions lesson and the capstone want a throwaway copy, because they create a workflow, a secret and an API key that you then retire.
The lessons
In order, each one building on the last; the capstone uses all seven.
Parallel sessions with worktrees
One developer, one repository, two sessions working at once — without either seeing the other’s half-finished edits. Git worktrees are the mechanism, and Claude Code has a flag for them.
Choose an orchestration pattern
Four ways to split a job across workers, and the differences that decide between them. The rule is ours, and it is short: pick the cheapest pattern the job needs.
Run and save a dynamic workflow
When the plan becomes code, three things follow: you can read it before it runs, independent agents can check each other’s findings, and a run that worked becomes a command you rerun on new input.
Script Claude Code with claude -p
Everything so far had you at the keyboard. claude -p is the same engine with the interface removed — one prompt in, one answer out, one exit code. Scripts and CI start here.
Claude Code in GitHub Actions
The same claude -p discipline, lifted into CI: a workflow file, a secret instead of a login, and every limit written down where no one has to remember it. Practised in a copy you throw away.
Share your setup with a team
CLAUDE.md was one developer’s memory. A team setup is configuration management: what everyone gets is committed, what is personal stays out, and the plugin is the unit of sharing.
Put bounds on autonomous runs
An unattended run does whatever you allowed in advance — so allow little and cap the rest. This is the lesson our whole thesis runs through: autonomy is a budget you set, not a switch you flip.
Capstone: CI review your team can share
No walkthrough this time — the brief says what the team needs, and you choose how to build it. Everything the seven lessons taught, in one pipeline, inside limits you can show your security team.
The Advanced exit
The capstone checks what files, git and CI can see; you tick the rest, honestly. Every line below is the level in one sentence — and the reason to trust an agent with work you do not watch.
Sources and attribution
This installment is our adaptation of Wesam Abousaid’s open curriculum Claude Code: Everything You Need to Know (v2026.10), used under its MIT license. The lesson spine, the exercise pattern and the capstone design are his; the phrasing, the opinions and the field notes from daily agent delivery are ours. Facts are stated as the official Claude Code documentation states them at the verified version above.
We are not affiliated with or endorsed by Anthropic. If you want the automated per-lesson checks and the purpose-built practice repo, use the upstream template.
- The course index — Beginner, and the three-installment roadmap
- Claude Code: Everything You Need to Know — the upstream repo ↗
- Run agents in parallel — Anthropic’s comparison of the patterns ↗
- The AI-native SDLC playbook — Claude Academy’s team course ↗
- All our guides — certifications, commerce, and this course