manastalukdar/ai-devstudio/prototype

Prototype

A prototype is throwaway code that answers a question. The question decides the shape.

Pick a branch

Identify which question is being answered, using the user's prompt, the surrounding code, or by asking:

• "Does this logic / state model feel right?" → Build a single shareable HTML file (free-play buttons plus tabbed guided walkthroughs) that pushes the state machine through cases hard to reason about on paper, driveable by a non-developer.
• "What should this look like?" → Generate several radically different UI variations on a single route, switchable via a URL search param and a floating bottom bar.

If the question is genuinely ambiguous and the user isn't reachable, default to whichever branch better matches the surrounding code (backend module → logic; page or component → UI) and state the assumption.

Rules that apply to both branches

1. Throwaway from day one, clearly marked as such. Locate prototype code close to where it will actually be used (next to the module or page it's prototyping for) but name it so a casual reader sees it's a prototype. For throwaway UI routes, obey the project's existing routing convention.

1. Trivial to run. A UI prototype starts from one command in the project's task runner: pnpm <name>, python <path>, bun <path>. A logic demo is a single HTML file the user double-clicks. No thinking required to start it.

1. No persistence by default. State lives in memory. If the question explicitly involves a database, hit a scratch DB or local file with a clear "PROTOTYPE, wipe me" name.

1. Skip the polish. No tests, no error handling beyond what makes the prototype runnable, no abstractions. The point is to learn something fast.

1. Surface the state. After every action (logic) or on every variant switch (UI), print or render the full relevant state so the user can see what changed.

1. Capture it when done. Fold any validated decision into the real code, then capture the prototype as a primary source: commit it to a throwaway branch out of main, and leave a context pointer to that branch on the implementation issue. Capture the answer (the verdict and the question it settled) in the issue or a commit. The main branch keeps only the validated decision.

Token Optimization

Expected range: 300–1,500 tokens (reads relevant files to understand existing conventions, then generates artifact)

Patterns used: Grep-before-Read (routing conventions, task runner scripts), early exit to branch selection, minimal file generation

Ported from [mattpocock/skills](https://github.com/mattpocock/skills) with attribution.