Skip to content

examples.md

Latest commit

 

History

History
84 lines (64 loc) · 3.15 KB

examples.md

File metadata and controls

84 lines (64 loc) · 3.15 KB
Raw

Examples

These are intentionally small. They exist to show how the framework should feel in practice.

Example 1 — App lane

Prompt

Build me a lightweight hiring workspace for startup founders to review applicants, track interview stages, and leave team notes. It should feel closer to Linear and Notion than a generic AI dashboard. Constraints: web, polished but not overbuilt, mobile decent, auth required.

Likely flow

  • lane: App
  • repo-first search for strong full-stack app / SaaS bases
  • choose the best auth/data/dashboard base
  • maintainability gate on routes, data access, naming, and state seams
  • polish dashboard hierarchy and candidate flow UI
  • verify core flow: sign in -> create role -> review candidate -> update stage

Receipt shape

  • chosen base: ...
  • why: ...
  • changed files: ...
  • commands run: ...
  • proof: ...

Example 2 — Dashboard / internal tool lane

Prompt

Build me an internal support dashboard for managers to see ticket backlog, blocked items, agent load, and SLA risk. It should feel clean and operational, not flashy. Constraints: desktop-first web app, mock data first, room for Zendesk later.

Likely flow

  • lane: Dashboard / internal tool
  • repo-first search for strong dashboard/admin bases
  • pick clarity and maintainability over bells and whistles
  • enforce readable data components and obvious state handling
  • finish with strong table/filter/status patterns
  • verify key flow: open dashboard -> filter backlog -> inspect blocked queue

Example 3 — Landing page lane

Prompt

Build me a launch page for an AI inbox copilot for founders. It should feel premium, fast, and direct — closer to Stripe or Arc than a typical glossy AI site. Constraints: one main CTA, mobile strong, restrained animation.

Likely flow

  • lane: Landing page
  • repo-first search for strong landing page / marketing bases
  • keep codebase smaller and avoid over-abstracting sections
  • focus Tier 3 on hierarchy, spacing, CTA emphasis, and purposeful motion
  • verify mobile rendering, CTA visibility, and page performance sanity

What these examples are proving

The toolkit is not trying to replace product thinking. It is giving a repeatable build discipline that improves:

  • base selection
  • maintainability
  • finish quality
  • verification

Example 4 — Report / Document lane

Prompt

Create an executive QA report for a founder deciding whether a new AI-built landing page is ready to share. Use the local test logs, screenshots, browser QA output, and known compromises. Tone: direct, decision-ready, non-technical. Format: Markdown with a source appendix.

Likely flow

  • lane: Report / Document
  • source-first intake: collect logs, screenshots, commands, and open issues
  • choose report skeleton before prose
  • separate facts, assumptions, and recommendations
  • format for skimming: summary, findings table, risks, next actions
  • verify links/files and render Markdown tables cleanly

Receipt shape

  • report audience: founder
  • decision supported: ready to share / hold / fix first
  • sources used: logs, screenshots, QA output
  • checks run: Markdown/table/link review
  • known gaps: unverifiable claims or missing sources