Cate Quickstart
Cate delivers a virtual software team in a box — built on the tools you already use: GitHub, Jira, and Linear. Spec agents that write tight plans, two-phase code reviews, and a full agentic SDLC produce quality code with context that’s visible to humans and agents alike.
Install and connect
Section titled “Install and connect”You’ll need:
- Claude Code or OpenCode installed
- Node.js >= 18
Download Cate, launch it, and activate your license.
Open a template project
Section titled “Open a template project”Start with a template so there’s nothing at stake. Choose Cate Quickstart — a one-page ShopCart app with nerdy products.
—TEMPLATE-PICKER-WITH-SHOPCART-HIGHLIGHTED—
ShopCart has three problems:
- Quantity stepper doesn’t work — clicking + and − does nothing
- Remove from cart isn’t hooked up — the button is dead
- No shipping options — the UI is missing entirely
Let’s fix them with Cate.
Plan the work
Section titled “Plan the work”Good agentic code starts with planning, not code. Click Build to open a planning session. When the agent asks if you want to build a new feature or fix a bug, choose new feature, submit, then copy and paste this:
ShopCart has three features to finish:1. The quantity stepper (+ and −) doesn't update the item count or totals2. The "Remove from cart" button doesn't do anything3. There's no way to select a shipping option
Use an epic with three parallel issues and get the work queued up.These are simple and for Cate's quickstart: skip spec review and sub-issue preview.If the agent’s unsure of itself, you’ll be asked to review a spec — if it looks good, tell the agent to go ahead.
—ISSUES-CREATED-CONFIRMATION—
Watch your agents work
Section titled “Watch your agents work”Autopilot is on, so code agents will pick up the issues automatically. Click any agent card to watch it edit code and run tests in real time.
—AGENT-CARD-LIVE-CODE-EDITING—
Check Home to see your virtual team, their work queue, and issues flowing through the pipeline.
—DASHBOARD-OVERVIEW-WITH-STEPPER—
While you’re there, an agent may need permission to run a task. You’ll see it alert you in the tab list, activity feed, or even with an OS notification if you’re doing something else. No more stuck bots.
Add review capacity
Section titled “Add review capacity”You’ve got three issues in flight but only two review agents — click the + Agent button on the AI Review row (or column, if you switched to board view) to add another so all three reviews can run in parallel.
Click into an active review and toggle Show reasoning to see the reviewer trace its logic as it reads the diff and checks the spec.
—REVIEW-AGENT-SHOW-REASONING—
Ready for review
Section titled “Ready for review”Review agents sign off, and issues land in your queue. Click one to start a review session — Cate summarizes the PR and related issues so you don’t have to read every diff.
—YOUR-REVIEW-SESSION-WITH-SUMMARY—
Run the app
Section titled “Run the app”ShopCart comes with a run config already set up. Wait for the human review agent to check out the branch for the PR, then click the green arrow beside Run the site at the top right. Run configs are one-click launch commands that start your app inside the agent’s worktree, so you’re seeing exactly what the PR delivers. Browse to localhost:3000 and test the three fixes: change a quantity, remove an item, select shipping.
—SHOPCART-RUNNING-WITH-FIXES—
Collaborate live
Section titled “Collaborate live”See “Free shipping on all returns”? Tell the agent to remove it — right from the review session. Then tell it to commit, merge the PR, and move the issue to Done.
Next steps
Section titled “Next steps”Keep experimenting with ShopCart:
- When the agent asks if you have follow-up work, try something: add a new product to the catalog, or plan something ambitious like payment integration.
- Launch a Code session to pair directly with an agent in real time.
Bring your own project:
- Open File → New Window and point Cate at any local git repo, or start an all-new project and it’ll set up a Git repo for you.
Go deeper:
- Plan & Delegate — spec features and investigate bugs
- Code — pair with an agent live
- Review & Merge — two-phase review in detail
- Run Your App — one-click launch commands
- Guardrails — quality gates agents must pass before PRs