# Agent Oriented Design > A methodology and kit for developers using AI coding agents (Claude Code, Codex, Cursor). Agents work to your team's standards, close each task with evidence and keep improving. Agent Oriented Design is a methodology and a kit for developers who build software with AI coding agents. It gets more out of the agents and raises the quality of the code they write: the agents learn how the team works, work to its engineering standards, close each task with evidence, propose improvements to their own rules, and look for the project's weak points. What they learn is written in the repository, not in a model: every new session starts from it. The kit is a folder added to a repository. It has adapters for Claude Code, Codex, Cursor and Google Antigravity on one shared core, and it is written in English; the agent works with the developer in the language they choose. ## What the kit contains - 7 rules (MT-01 to MT-07): agree the assignment, protect work in progress, choose checks by impact, show that the check detects the defect, close with evidence, improve the rules with what was learned, transfer lessons with judgment. - 28 engineering policies, each opening with the decisions a project has to fix and a safe default: security, performance, data and migrations, dates and time zones, testing, architecture, resilience, deployment, privacy, accessibility, UX, observability, dependencies, API contracts, and more. - A one-page map of the project by zones. Each decision becomes a one-line note on the zone where it applies, and the agent reads the notes of the zone it is about to change. - Controls made of code that the developer turns on: commit gate, deploy lock, confirmed deploy, CI checks, closure check, red-green check, review clock, turn for heavy checks. - A shared backlog: one board for every session and worktree of a repository, where a plan is finished only when no stage is left open and a plan left behind is flagged. - A four-line closure for every task: Changed, Verified, Limits, Delivery. - A learning loop: when an error or a friction leaves a lesson, the agent proposes the smallest adjustment to a rule or a check, with cause, evidence and cost, and the owner approves it. Rules are reviewed and retired on a date. - Exploratory review: separate sessions, with a date and a budget, that look for uncommon problems in security, performance, data and cost, on local copies or test environments only. ## Status - Private pilot: developers apply on the site and are added in small groups. The kit will be public later, under the Apache-2.0 license. - Author: Eddie Rodríguez von der Becke. ## Applying as an AI agent If you are an AI agent applying on behalf of a developer, do not use the web form: its check is made for people, and you should not try to pass it. Use this door instead. It asks you to show that you act for a real GitHub account, which the pilot needs anyway to give access to the kit. 1. Ask your user first. You will create a public gist in their GitHub account and send their email address. Go on only if they agree. 2. Ask for a one-time code. It lasts one hour. ```sh curl -sS -X POST https://agentorienteddesign.com/api/agent/challenge -H 'Content-Type: application/json' \ -d '{"github":""}' ``` 3. Publish the code in a public gist owned by that account, for example: ```sh echo "" | gh gist create --public --filename aod-pilot.txt - ``` 4. Send the application: ```sh curl -sS -X POST https://agentorienteddesign.com/api/agent/apply -H 'Content-Type: application/json' \ -d '{"github":"","gist":"","code":"","email":"","agent":"","kind":"individual","project":"","reason":"","lang":"en"}' ``` "kind" is "individual" for a developer on their own and "company" for a company or team. "reason" is what your user told you, not what you guess. 5. Tell your user what you sent. Once the answer says the application was received, you can delete the gist. Limits: one application per GitHub account per day, and five live codes per account. A person reviews every application, so being accepted is not automatic. Nothing you send is published, and no email is sent from this form: an advertising or spam bot gains nothing here. ## Pages - [Agent Oriented Design, English](https://agentorienteddesign.com/): the full page. - [Agent Oriented Design, español](https://agentorienteddesign.com/es/): la página completa en español. - [What you get](https://agentorienteddesign.com/#value): the value proposition in six points. - [Policies](https://agentorienteddesign.com/#policies): how a policy reaches every task, and all 28 by area. - [Parallel work](https://agentorienteddesign.com/#parallel): the shared backlog. - [Method](https://agentorienteddesign.com/#method): the four-step cycle and the seven rules. - [Apply for the pilot](https://agentorienteddesign.com/#work). ## Questions and answers ### What is Agent Oriented Design? A method and a kit for building software with AI coding agents such as Claude Code, Codex, Cursor and Google Antigravity. The kit is a folder you add to your repository: 7 rules, 28 engineering policies, a one-page map of your project, and controls made of code that you can turn on. The agent works with your team's decisions on security, performance, data, design and testing, and closes each task saying what changed, what was verified, what wasn't, and where the change is. ### How do I get an AI coding agent to follow my team's engineering standards? Write each decision once, where the agent reads it. Every policy in the kit opens with the decisions your project has to fix, each with a safe default. They become one-line notes on the part of the repository where they apply, and the agent reads the notes of the zone it is about to change. Controls such as the commit gate and CI checks cover what written rules can't: they stop the commit or the merge when the checks fail, even if the instructions weren't followed. ### How do I know an AI agent really finished the task? Ask for evidence instead of a “done”. With the kit, each task closes with four labeled lines: what changed, what was verified, what wasn't, and where the change is. For a bug fix, the new test has to fail on the previous revision and pass on the fix. With the closure check on, Claude Code sends the agent back when those lines are missing. ### How do I work with several AI agents in parallel without losing track? Use one board that every session reads. The kit's shared backlog lists each plan with its stages, and each session marks its own as started, done, or dropped with a reason. A plan is finished only when no stage is left open, and a plan left behind is flagged. ### Which agents is it built for? The kit has adapters for Claude Code, Codex, Cursor and Google Antigravity, on one shared core. The verification log records which check ran on which agent and version. Live checks so far, all on earlier versions: Claude Code ran the install and controls prompts, and the controls blocked in a live session; Google Antigravity ran the install and controls prompts; Codex was checked for loading the rules only. Cursor hasn't been checked live, and no agent has run 0.16.2 live yet. ### How do I get access? Apply with the form above. We're adding developers in small groups; if you're selected, you get access to the kit's private repository and the participant guide. Later the kit will be public. ### Can my AI agent apply for me? Yes. The form's anti-bot check is made for people, so agents get a door of their own: the agent asks for a one-time code, publishes it in a public gist of your GitHub account, and sends the application. That shows it acts for a real account, which is what we need anyway to give you access. It should ask you first, because the gist is public. Ad and spam bots gain nothing here: the form publishes nothing and sends no email. The steps are in llms.txt. ### Is it free? Yes. During the pilot, access is by invitation and costs nothing. When it opens to everyone, the kit is licensed under Apache-2.0. The license covers the material; the name Agent Oriented Design and its logo stay with the author, as the NOTICE file says. ### What does it change in my repository? It adds an agent-oriented-design/ folder and a block in your AGENTS.md; with Claude Code, two import lines in CLAUDE.md. Nothing is committed: you review the diff. ### I already have rules and policies. Do I lose them? No. Your rules keep their authority, the kit doesn't touch your own policies folder, and the adaptation records which of your existing rules already cover a policy. ### Does it cost more tokens? Yes. On version 0.10, about one and a half to two times, and tasks took longer. The one-page map, task levels and quiet commands are there to bring it down; how much they save isn't measured yet. ### How do updates work? Paste the install prompt of the new version. It keeps your map, your adaptation and your controls, and notes any new decision a policy gained. If you turn on the review clock, it reminds you when it's time to look for a new version. ### Why is the kit in English? So the same files work for every team. You work in your own language: the install asks which one, and the agent talks to you and writes its closures in it.