factory

A dark factory that runs on your Mac: Linear tickets in, merged GitHub PRs out, with almost no human in the loop.

Every 10 minutes a small Node script looks at your Linear tickets, your GitHub PRs and your running Claude Code sessions, then starts or wakes a Claude session to move each ticket one step forward. The script itself makes no model calls, and Linear is the only state of record.

Backlog ──triage──► Todo ──build──► In Progress ──PR──► In Review ──merge──► Done
                                         │
                                         └── stuck? one question on the ticket, you reply, it resumes

What it uses

Piece	What it does
Claude Code	every worker is a background claude --bg session in its own git worktree
Linear	tickets, status and the factory's state (one machine-owned comment per ticket)
GitHub + gh	PRs, CI, reviews, merges (gh stack for PR stacks)
macOS launchd	runs the loop every 10 minutes and keeps the dashboard up at http://localhost:7788
UI Verify	the visual check: the agent sees what its UI change did and fixes it before a human looks

The skills

All in skills/, linked into ~/.claude/skills/ by setup.sh.

The loop (the same skills as UI Verify's uiverify-workflow plugin):

Skill	What it does
factory	one change from ask to merge-ready PR: design if needed, implement, review, verify, open the PR
design	drafts a design and has two models critique it until it holds up
review-loop, review-multi-model	Claude and Codex review the diff, verified fixes get applied, repeat until clean
e2e-verify	drives the change through the real app with Playwright
babysit-pr	watches CI and review comments on a PR and fixes what fails
add-rule	turns a correction into a CLAUDE.md rule plus a lint guard, so it never happens again
evaluate-run	grades a finished run against your conventions

The factory (what runs it unattended):

Skill	What it does
factory-auto	factory with no questions; also the babysit pass that fixes, decides and merges
factory-triage	reads a Backlog ticket: ready, split, needs a decision, or later
factory-plan	groups ready tickets into packages, one PR stack each
factory-project	the one place you spend real time: spec, architecture and the ticket tree, together
factory-run	start, stop, pause, status, in plain words
factory-monitor	read-only health check every 30 minutes, tells you when the loop quietly stalls

For visual testing, add UI Verify's own skills: npx skills add uiverify/uiverify.

Install

git clone https://github.com/uiverify/factory ~/.claude/factory
cd ~/.claude/factory && sh bin/setup.sh

Then put your Linear API key in .env, set your repo, Linear team and one target in factory.json, and fill in principles/principles.md. That last file is the important one: it is the judgment workers use instead of asking you (what to decide alone, when to stop, when a PR may merge). The template has examples.

Try one tick without side effects, then start it:

node bin/factory-tick.mjs --dry-run --verbose
sh bin/install-launchd.sh 600        # or /factory-run start in Claude Code

Start small: one target, one or two tickets in flight, and every merge waiting for a human approval.

Before you run it

• This is one person's working setup, published as is. Not a polished tool. Expect to read the code and adapt it.
• Workers run with no permission prompts (--dangerously-skip-permissions). A worker can run any command as you, with your credentials. Run it only on a machine and accounts where that is fine.
• macOS, Claude Code, Linear and GitHub only.
• It is only as good as your repo. Conventions in CLAUDE.md, lint and tests that catch mistakes, a way to see UI changes. The factory runs the loop; your repo supplies the judgment.
• It uses your Claude subscription. Several Opus sessions at once is heavy.

How it works in detail, every config key, and the known quirks: docs/how-it-works.md.

MIT licensed.