Live LabsComing soon10 min

Run Claude Code as a Background Agent

Run Claude Code from a GitHub workflow, so a labelled issue becomes a pull request while you do something else.

Out on Wednesday 14 October10 minutes, one build start to finish

Get it in your inbox the day it is out

One email when the recording is live, then the newsletter. Unsubscribe in one click.

We email you each new issue. You can leave at any time. Privacy

Why this matters

Claude Code in a terminal still needs you at the keyboard. Running it from a GitHub workflow turns a labelled issue into a pull request while you work on something else, and keeps the review step where it already is: in the pull request.

What you build

Claude Code picking up an issue on its own and opening a pull request for you to review.

  1. Install the Claude GitHub app and add your API key as a repository secret.
  2. Add the workflow that starts Claude Code from an issue.
  3. Trigger it on an issue, and Claude Code works on it in a GitHub Actions runner.
  4. Review the pull request it opens like any other.

The ideas behind it

  • Triggered by an issue

    A label or comment starts the run, so anyone on the team can hand work to the agent.

  • Runs in CI

    Claude Code works in a GitHub Actions runner with the repository checked out, not on your machine.

  • Reviewed like any change

    The result arrives as a pull request with its diff, tests and checks, and goes through your normal review.

Before you start

You need
A GitHub repository you can add workflows to, and an Anthropic API key.
Length
About 10 minutes, one build start to finish.

Questions

Is it safe to give an agent repository access?

Its access is limited to what the workflow and app allow, and nothing merges without your review.

What does it cost?

Your Anthropic API usage plus normal GitHub Actions minutes.

Does it work on private repositories?

Yes, with the GitHub app installed on that repository.

Is the recorded lab free?

Yes. Watch it any time, follow the steps and keep the code you build.

How is it different from a Live Lab?

A recorded lab is one short build you follow on your own. A Live Lab builds a bigger part of the system in an hour, with time for your questions.

Then take it further

This build is one piece. The Live Lab Build an AI Mission Control App goes from here to a working part of a real system.

Build the board
Backlog, Needs you, In flight and Done, with each agent's role and tools.
Ask a human first
An agent pauses and asks on Telegram before an action it cannot undo.
Map it to your company
Swap the connectors for the tools your team already uses.

The parts of a production agent system that lab builds, lit on the map:

Week 1Architect
  • Roles and modelsOne job per agent, a model chosen for that job, and a token budget.
  • OrchestrationA queue agents pick work from, with hand-offs a person can follow.
  • Context and memoryRetrieval with sources, memory across sessions, prompts laid out for the cache.
  • Tool callingTyped tools that act in GitHub, Linear and Google.
Week 2Build
  • MCP serversEvery tool behind an MCP gateway, scoped to the role that needs it.
  • Durable executionRuns that resume after a crash and never repeat a write.
  • Human in the loopA person approves anything that cannot be undone.
Week 3Secure and deploy
  • Identity and permissionsEach agent signs in as itself and acts on behalf of a user.
  • LLM gatewayRouting, caching and a budget on every model call.
  • Traces, evals and costEvery run traced, scored and charged to the agent that made it.
  • Multi-tenancyEach team or customer kept apart, in data and in the bill.
Week 4Govern and extend
  • Policy as codeRules checked on every action, not written in a document.
  • Signed skillsNew roles built from reviewed, signed skills.

More recorded labs