Skip to main content

How I build things with AI agents

My main project is studio-kit. It runs studios of AI agents that build games, videos and this site. It changes quickly, so this page has two tabs: how it runs now, and how it ran on 24 October 2026.

01 · Since 8 Nov 2026

studio-kit: how the studios run now

Since 8 Nov the studios run without ticks. I play the client: I set a goal, the studio drafts the features, and I check each one when it's finished. Agents and scripts do everything in between.

The studio Viewer's All studios screen on 13 Nov 2026: six studios, their builders and merges, and the day's filed slices.

How a change gets in

  1. 1 Brief The designer writes one brief per feature and cuts it into 2-3 large slices.
  2. 2 Slices A builder claims a slice from the work queue and builds it end to end in its own git clone.
  3. 3 Merge queue A script, not an AI, lands each slice on main as one commit when the full check is green.
  4. 4 Review after merge A risky merge gets a reviewer session after it lands. A finding goes to the top of the queue as a bug slice.

Nothing else gates a merge. A slice that fails the check twice goes to the lead.

Five roles

  • Designer Writes each feature's brief: the layout, the copy, the slices and a checklist for me.
  • Builder Builds one slice end to end, in its own clone.
  • Artist Builds the art slices.
  • Lead Turns triggers into actions, such as my notes into bug slices. Writes no code.
  • Reviewer Reads risky merges after they land. A severe finding becomes a revert slice.

Builders come from Claude and Codex in most studios. Rich to Richer still runs Claude, Codex and Gemini together.

No human gates

I check finished features from a checklist on the studio Viewer's New to play page, and send notes. Within one lead session each note becomes a bug slice or an answer, or it is declined with a reason. Open choices don't wait for me either: the designer picks a default and lists it as 'I chose X; reply to change it'. Only real-world actions wait for me: a deploy, money and accounts.

What changed since 24 October

24 October Now
Ticks Features and slices
Seated cast Five roles
Review before merge Review after merge
QA gate The client's checklist
Forum and board Notes become bugs
52 rules 12 rules

Still true: developer runs start with secrets stripped, each builder works in its own git clone, every request goes on the cost ledger, and I still play the client. The 12 rules come with caps (CLAUDE.md at most 8 KB, each role brief at most 4 KB) that a kit test enforces.

  • 745 slices landed by the merge queue since 8 Nov, across four studios
  • 12 rules in the kit's rulebook, down from 52
  • 2,440 kit tests: 1,490 Vitest and 950 node
  • 90 of the kit's 359 commits came after 8 Nov; it is at version 0.7.1

As of 13 Nov 2026, from each studio's work state and the kit's git history. The slice count is approximate: Reel Engine 477, Rich to Richer 150, SuperMarketGame (Corner to Chain) 111, this site 7.

Short on time? Ask your AI.

Paste this into Claude, ChatGPT or any assistant that can read the web. It reads a plain-text brief written for models and follows the links.

Read the brief yourself

Read https://connorjarvis.co.uk/llms.txt and the pages it links to. Then tell me what Connor Jarvis has built with AI agents, the engineering decisions behind it, and the evidence for each claim. Say plainly what you couldn't verify.

What the studios make

The kit started with games. It now has a video kind too: Reel Engine, and the What Do You Do? series. This site is a studio as well; this page was built as a feature in slices.

See what the studios made

02 · Feb–Oct 2026

Earlier AI work

  1. 17–18 Oct 2026

    Frostfall TD

    A maze tower-defence game a Claude agent built on its own from a written spec, balanced with 450 bot-played games.

  2. Mar–Sep 2026

    Covered

    Insurance comparison by conversation, using Claude Haiku with streaming tool use and a rule-based check for wording that could read as regulated advice.

  3. Feb–Sep 2026

    Spectre

    A paper-trading signal scanner in which Claude analyses the top candidates, with a cost estimate before every call and a daily budget.

  4. Mar–Jul 2026

    Inference

    An idle game where you are the language model, designed through rounds of designer, tester and storyteller agents and balanced by simulation.

  5. Jul 2026

    interview-me

    Agent skills for Claude Code, Copilot CLI and Codex: one interviews you on the decisions in a change you've shipped and scores you, the other reviews a branch and drafts the comments.

  6. Jul 2026

    Azure DevOps skills

    Two agent skills: one digests sprint health, PRs and builds into an HTML report, the other is a thinking partner for a single work item.

  7. Jun 2026

    human-voice

    A Claude Code skill that makes the prose an agent writes for you read like you wrote it.

  8. 15–30 Mar 2026

    Asset Forge

    An MCP server with 20 tools that lets Claude Code generate game art and audio through ComfyUI on my own GPU; 102 of its assets went into Space Game and Inference.

  9. 3–25 Mar 2026

    Space Game

    A space strategy game in Godot whose difficulty was tuned from 273 automated playtest games.

Most of my other projects were built with Claude Code too, without AI in the product. See the projects

03 · Mar 2025 – now · Volunteer

Somerton RFC, live for a real club

I redesigned and built Somerton RFC's website, and I run it for the club for free. A scraper reads England Rugby's system overnight, so fixtures, results and league tables stay current with nobody typing them in, and it alerts my phone when a run fails. 177 of its 346 commits were written with Claude Code.

Visit somertonrfc.co.uk
The Somerton Saxons homepage: a match photo behind the club name, with the latest result card and a Join the club button.

04 · 2026 · Evenings and weekends

The night shift

I help run a small game-server hosting company, but only in the evenings and at weekends. The agents cover the hours I'm at my day job or asleep. The code is the company's, so this is described rather than linked.

  • 166 MCP tools, co-built, so staff can run the hosting control panel through Claude
  • 6 headless Claude Code pipelines I wrote, each on a schedule
  • 3 fact-checking agents from my Claude plugins that review article drafts in parallel
  • 200+ isolated git worktrees, so coding agents work side by side without colliding

The rota

  1. Every 15 min Support triage Haiku picks up new tickets, an agent investigates each one through the MCP tools, and drafts a reply for a person to review. Spend is capped per run.
  2. 01:30 Competitors Checks which games rival hosts sell that we don't, and flags any they've just added.
  3. 08:00 Game news A daily digest, with a second AI score for SEO potential.
  4. 08:30 Company tasks What's overdue or due today, ranked by AI, with a two-week look-ahead.
  5. 09:00 & 21:00 Reddit Hosting-related posts matched across 60+ game subreddits all day, summed up in an AI-scored digest.
  6. Weekly Partner scouting Creator prospects collected and scored by AI.