| name | mature-app-starter |
| description | Use when a greenfield app needs a mature foundation plan before implementation. |
Mature App Starter
Use this skill to establish a strong baseline before building a new app. Keep the output practical:
decisions, tradeoffs, file structure, verification, and next actions.
First Pass
Start by identifying:
- Target users and their high-stakes moments.
- Core job-to-be-done and first valuable workflow.
- Platform target: web, mobile, Expo, native, desktop, or API-first.
- Domain risk: privacy, safety, compliance, accessibility, payments, identity, user-generated content, regulated data, or sensitive workflows.
- Current stage: idea, prototype, MVP, beta, production, or rebuild.
If the user has not provided enough context, make conservative assumptions and label them.
Product And UX Baseline
- Build the real primary workflow first, not a landing page.
- Keep information architecture simple: primary nav, main task surface, settings/account, and support/legal when needed.
- Design for repeated use: dense enough to scan, clear states, predictable navigation, and low-friction recovery.
- Include empty, loading, error, offline, permission-denied, and success states early.
- Prefer accessibility by default: semantic controls, readable contrast, scalable text, keyboard/screen-reader paths, and touch targets.
- For sensitive domains: avoid overclaiming, protect user data, clarify escalation paths where relevant, and design for trust and auditability.
Technology Baseline
- Prefer boring, well-supported technology unless the product need justifies complexity.
- Choose one app architecture and document it early.
- Keep route/screen files thin; put product logic in feature modules.
- Keep external systems behind typed service boundaries.
- Centralize environment config and never scatter secret or environment reads across UI code.
- Use structured domain types close to the feature first; promote to shared types only after reuse is real.
- Add observability, analytics, and error reporting behind wrappers before calls spread.
Engineering Setup
Recommend a starter structure like:
app-or-routes/
src/
features/
components/
services/
lib/
config/
types/
test/
docs/
adr/
project-index.md
engineering-workflow.md
Create or recommend:
- Root
AGENTS.md for project-wide rules.
- Domain
AGENTS.md files for src/features/, src/services/, and backend/schema folders when they exist.
- ADRs for stack choice, backend strategy, testing strategy, analytics/observability, and sensitive-data handling.
- A focused task workflow skill only when the project has a repeated task process.
Verification Baseline
- Typecheck and lint from the start.
- Unit-test domain logic, calculations, parsing, permissions, and state transitions.
- Component-test critical UI states where practical.
- Add end-to-end smoke tests after the first stable workflow exists.
- Manual UX checks are required for responsive layout, accessibility, and mobile ergonomics.
Security And Data Baseline
- Classify data by sensitivity before modeling storage.
- Keep secrets server-side and out of app bundles.
- Add auth only when the product workflow needs identity; when added, design roles and recovery paths deliberately.
- For sensitive or regulated domains, treat privacy, audit trails, consent, retention, and export/deletion as first-class requirements.
- Avoid collecting data that is not needed for the first useful workflow.
Blueprint Reference
For deeper guidance, read references/index.md first and then load only the specific referenced
file needed for the user's request.
Output Shape
For a new app plan, return:
- Recommended product scope.
- Suggested architecture and stack.
- Starter folder structure.
- UX principles and key states.
- Security/privacy assumptions.
- Testing and verification plan.
- Suggested
AGENTS.md layers.
- Next 3 to 5 implementation tasks.
If implementation begins, keep changes small and create the foundation before feature breadth.