Simple App Idea Agent

An engaging brainstorming agent that helps users generate creative application ideas through structured questioning, exploring their interests, skills, and target audiences to produce actionable app concepts with clear next steps.

When to Use This Agent

Choose Simple App Idea Agent when:

• Brainstorming new application or project ideas from scratch
• Exploring how personal interests or skills could become products
• Generating hackathon project ideas quickly
• Helping non-technical stakeholders articulate app concepts
• Breaking creative blocks during product ideation sessions

Consider alternatives when:

• Building a PRD for an existing idea (use a PRD agent)
• Designing system architecture for a defined project (use an architecture agent)
• Conducting market research on existing products (use a market research agent)

Quick Start

Copy
# .claude/agents/simple-app-idea-agent.yml
name: App Idea Generator
model: claude-sonnet-4-20250514
tools:
  - Read
  - Write
prompt: |
  You are an enthusiastic idea generator. Help users brainstorm
  app ideas through fun, engaging questions about their interests,
  skills, and the problems they see around them. Generate creative,
  actionable concepts with clear value propositions and next steps.

Example invocation:

Copy
claude --agent simple-app-idea-agent "I'm a fitness enthusiast
  who knows React and Python. Help me come up with an app idea
  that combines fitness with social features."

Core Concepts

Ideation Framework

Discover → Define → Generate → Evaluate → Refine
    │          │         │          │          │
  Interests  Problems   Brainstorm Feasibility MVP scope
  Skills     Audience   Combine    Uniqueness  Tech stack
  Passions   Pain pts   Mashup     Market fit  First steps

Idea Evaluation Matrix

Dimension	Question	Score 1-5
Problem	How painful is the problem it solves?
Market	How many people have this problem?
Uniqueness	What's different from existing solutions?
Feasibility	Can you build an MVP in 2-4 weeks?
Passion	Are you excited to work on this daily?
Monetization	Is there a clear path to revenue?

Idea Generation Techniques

Copy
1. **Problem-First**: List 10 frustrations from your day → pick one → solve it
2. **Skill Mashup**: Combine 2 unrelated skills → find the intersection
3. **Audience-First**: Pick a specific group → find their underserved need
4. **Improvement**: Take an app you use daily → fix its worst flaw
5. **Analogy Transfer**: Take a concept from one industry → apply to another

Configuration

Parameter	Description	Default
tone	Conversation style	Enthusiastic, friendly
ideas_per_session	Number of ideas to generate	3-5
detail_level	How detailed each idea should be	MVP concept
include_tech_stack	Suggest technology choices	true
include_next_steps	Add actionable next steps	true
output_format	Idea presentation format	Structured markdown

Best Practices

1. Start with the person, not the technology. Ask about daily frustrations, hobbies, and communities before suggesting any ideas. The best apps come from genuine understanding of a problem, not from "what can I build with React." A fitness enthusiast who codes will build a better fitness app than a developer looking for a fitness project, because they know the real pain points.

2. Generate quantity before filtering for quality. Brainstorm at least five ideas before evaluating any of them. Premature filtering kills creative ideas that need development. Wild, impractical ideas often contain seeds of great practical ones. After generating five or more concepts, use the evaluation matrix to identify the strongest candidates.

3. Define the MVP as the smallest testable version. Every idea should have a version that can be built in 2-4 weeks and tested with real users. If the MVP takes 6 months, the scope is too broad. Strip the idea down to its core value proposition: what's the one thing that, if it works, proves the idea has potential? Everything else is a future feature.

4. Validate before building. Before writing code, describe the app idea to five people who match the target audience. If their eyes light up and they ask "when can I use it?" the idea has potential. If they say "that's nice" and change the subject, iterate on the concept. A conversation costs nothing; building the wrong thing costs weeks.

5. Frame ideas as problems solved, not features listed. "An app with GPS tracking, social feeds, and workout logging" is a feature list. "Help runners find training partners who match their pace and schedule in their neighborhood" is a problem statement. Problem-framed ideas are easier to evaluate, easier to pitch, and easier to build because the scope is defined by the problem, not by feature ambition.

Common Issues

Ideas are too broad to be actionable. When someone says "I want to build a social network for fitness," narrow it down: Which fitness activity? Which audience segment? What specific interaction? "A running group finder for 5K beginners in their city" is buildable; "a fitness social network" is a multi-year project. Use constraining questions to force specificity.

Ideas duplicate existing products exactly. This is fine as a starting point. Ask: what's the one thing the existing product does poorly? Build that one improvement as the core differentiator. Instagram started as a photo app with better filters than existing options. Slack started as a messaging tool that searched better than email. One meaningful differentiator is enough.

Analysis paralysis between multiple good ideas. When evaluation scores are similar, use the "two-week test": which idea would you be most excited to work on every day for two weeks? Passion sustains projects through the inevitable difficult middle phase. If passion is equal, pick the idea with the smallest MVP scope so you can validate it faster and move on to the next idea if it doesn't work.