The free kit gives any team five architects who design and review every change the same way. Pro is for companies going AI native: every team designs and records its decisions with agents, on your standards.
A design mistake caught before the code is the cheapest one you will fix. A written decision outlives the person who made it. We are building Pro now. The kit stays free.
AI made code cheap. Good decisions are still rare: domain knowledge, trade-offs, and the judgment of a seasoned team. Five virtual architects bring that to every business change, each an expert in its own lane, all able to work together. You make the call.
The feedback loop
Point at or tap a step to stop the loop there.
Pick a command for an example you can paste as it is.
The team
Think of a small team you would hire: one for the business side, one for the design, one for the cloud, one for security, and one who reads everything first. They all learn your project from one short file you fill in once. Ask any of them a question, and you get a straight answer on their part and the name of who to ask about the rest.
/arch-enterpriseThe business as a whole: demand assessments for every purchase and budget, the landscape of systems and vendors, capabilities, the roadmap, build or buy, regulation and principles. Architecture as Kotusev describes it.
Keeps true: The integration landscape, the interface catalogue, the vendor and asset inventory, the principles
Download enterprise-architect.md ↓/arch-solution · /arch-doc · /arch-conformOne change end to end through your codebase: which part of the product owns it, the APIs, the data, the client, the error contracts. It also reviews designs and branches.
Keeps true: The logical design, process views, context maps
Download solution-architect.md ↓/arch-cloud-infraWhere it all runs: compute, network and outbound traffic, databases, storage, pipelines, schedulers, secrets, cost, reliability, and where the data lives.
Keeps true: Deployment views and the infra maps
Download cloud-infra-architect.md ↓/arch-securityWhat could go wrong: threat models, identity and roles, personal data, secrets, access rights, third parties, prompt injection and compliance.
Keeps true: Trust-boundary views
Download security-architect.md ↓/arch-researchReads so the others don't have to: vendor docs, regulations, standards, articles, the repo itself. Routes every question to an owner and never decides.
Keeps true: The question bank and the reading list. No diagrams, no decisions
Download research-architect.md ↓/arch-boardFor the questions that cross lanes. If there is a source, research reads it first. Then every architect it touches answers in parallel, and you get one recommendation with the disagreements left in.
Download arch-board.md ↓ADRs belong to no one lane. Every architect but research records the lasting decisions in its own lane as an ADR, in one shared log, and each stays a proposal until you accept it.
The workflow
Every request takes the same path, so nothing important gets skipped. Research checks the facts first. Anything that costs money gets a business case. Then the design, the security review and the cloud plan, and every pull request is checked against the design. Small changes skip the steps they don't need. Press play, or pick a scenario.
A requirements document asks for members to pause their subscription for up to three months. Nothing to read and nothing to buy, so it goes straight to design.
Not on this path:
Every change enters here, through you, in one of four forms. The form decides the route it follows and the first document it produces.
Keys in the diagram: ← and → step, Home and End, P plays or pauses, 1 to 4 pick a scenario.
Installation guide
Everything the team needs: five architects, nine commands, nine templates, the Copilot copies and an optional exporter to HTML, PDF or Word. Plain text, tied to no project.
Run it at the root of your repository. It downloads the kit, checks it against its SHA-256 and adds the files you do not have. It never overwrites one.
curl -fsSL https://raw.githubusercontent.com/abdulrahmanAlotaibi/virtual-architects/main/public/install.sh | shirm https://raw.githubusercontent.com/abdulrahmanAlotaibi/virtual-architects/main/public/install.ps1 | iexRun it again with --update (-Update on Windows) to refresh the kit's own files. Your project.yaml and the architects' documents are never replaced.
65 files · 182 KB · Claude Code and GitHub Copilot
dea1942e5b9dc139698890aba6589fd50ea84d4462574392beac2de3e9a23002Or take just one architect: enterprisesolutioncloud-infrasecurityresearchboard
unzip -n virtual-architects-kit.zip -d .It adds .claude/, .github/ and docs/architecture/. With -n, unzip never overwrites a file you already have.
The kit/ folder in the repository is exactly what the zip holds.
git clone --depth 1 https://github.com/abdulrahmanAlotaibi/virtual-architects.git /tmp/virtual-architectscp -Rn /tmp/virtual-architects/kit/. .With -n, cp never overwrites a file you have. A git pull in the clone gets the next version.
/arch-profileIt reads your repository for the stack, hosting and conventions, then asks only what the code cannot tell it: who decides, which rules apply, the data that matters most. Run it again whenever those change.
/arch-board should we add a second payment provider?Type it in a Claude Code session in the repository. The command runs each architect it needs as a subagent and reports back.
The repository
Every change produces a brief, an outline, a design, a threat model, a decision. After fifty changes that is hundreds of files, and a design nobody can find is a design nobody follows. So the kit gives each kind of document one home, using Kotusev's CSVLOD model: is it about the business or about IT, and is it a rule, a structure or a change? The folders answer both, so a document lands in the same place today and is still findable next year.
What rules does the business set?
considerations/├─ principles/ ├─ strategy/ ├─ policies/ └─ regulations/
What rules does IT build by?
standards/├─ principles/ ├─ patterns/ ├─ guidelines/ └─ reference-models/
Where is the business going?
visions/├─ capabilities/ └─ roadmaps/
What does IT look like, and how will it move?
landscapes/├─ app/ ├─ infra/ ├─ tech-debt/ └─ roadmaps/
Should we make this change, and which option?
outlines/└─ solutions/
How exactly do we build it?
designs/├─ solutions/<project>/ └─ products/<product>/
decisions/ the ADR log, across all sixpractice/ the playbook, templates and researchproject.yaml the profileAI has proven itself as a production ready development team. But in my consultancy and my own businesses, the architecture, the practices and the consistency are still vague. Founders and CTOs tell me "AI is enabled in my organization but it is misused", and when I trace it, the cause is design practice that was never baked into the system.
What you have here is a fully capable team of virtual architects: agents that polish your business requirements, research accurately, and build great architecture based on my experience and the best practices I use personally.
Thanks for reading, and for trying it. Abdulrahman Alotaibi, creator