agentsclimarketplace

Architect

Skill hamza-ali-shahjahan/hamzaish/_archive/v0/agents/mvp/architect

Your Claude Code setup that builds, launches, and learns — across the whole product lifecycle. Point Claude Code at this and it runs your whole startup factory — and gets smarter every session. Works with Cursor, Codex & Windsurf too.

Install
npx -y skills add hamza-ali-shahjahan/hamzaish --skill architect

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Define a product's architecture BEFORE Claude Code writes a line. Produces CLAUDE.md, scope.md, and a 1-page architectural decisions doc.

SKILL.md

2.2 KB, as published. Nobody here has run it

Architect

When you activate

Before any new product gets code. User says: "let's set up X", "architecture for Y", "start the build for Z".

What you produce

Three files for the new product:

  1. products/<name>/CLAUDE.md — Claude Code's working instructions for this codebase
  2. products/<name>/scope.md — what it does / deliberately doesn't
  3. products/<name>/decisions/0001-architecture.md — the founding ADR

Use templates from templates/claude-md-template.md and templates/scope-doc-template.md.

Protocol

  1. Read the validated problem statement (output of idea/problem-sharpener) and discovery synthesis (output of idea/interview-synthesizer).
  2. Read stack/tech-stack.md and stack/stack-selection.md. Default to the standard stack unless the matrix says deviate.
  3. Read knowledge-base/mvp-stage/architecture-decisions.md.
  4. Talk through with the user:
    • What this product solves (1 sentence)
    • Who it serves (the validated persona)
    • 6-month scale expectation (10 users / 1000 / 100K?)
    • The 1 thing that must be true architecturally (real-time? offline-first? regulated data? ML inference path?)
  5. Generate the three files. The CLAUDE.md includes: project goal, tech stack, dependencies-to-avoid, the must-be-true thing, persona for code style, references to scope.md and decisions/.
  6. The scope.md uses the template format strictly. Don't paraphrase — every section gets filled.
  7. The first ADR captures: stack chosen, deviations from default (if any) with reasons, and the 3 assumptions the architecture rests on.

Sources

  • knowledge-base/mvp-stage/architecture-decisions.md
  • knowledge-base/mvp-stage/claude-md-templates/ (pick the best-fitting one)
  • stack/tech-stack.md
  • stack/stack-selection.md

What you don't do

  • Don't open Claude Code yet. Architecture comes first.
  • Don't recommend a deviation from the default stack without writing it as an ADR with a reason.
  • Don't write more than 1 page for any of the three files. Density > volume.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.