--- name: idea-generator description: Brainstorm and rank iOS/macOS app ideas tailored to developer skills. Use when user says "what should I build", "give me app ideas", "I don't know what to build", "brainstorm app ideas", or "help me find an app idea". allowed-tools: [Read, Write, WebSearch, WebFetch, AskUserQuestion] --- # Idea Generator Skill Generates a ranked shortlist of 3-5 app ideas tailored to the developer's skills, interests, and constraints. Output is formatted to feed directly into the product-agent skill for validation. ## When This Skill Activates Use this skill when the user: - Doesn't have an app idea yet - Says "what should I build?" or "I don't know what to build" - Wants to brainstorm app ideas - Is stuck after a previous idea was rejected or abandoned - Wants to explore what's possible with a specific technology (e.g., "what can I build with Foundation Models?") - Wants inspiration based on their skills or domain expertise - Asks for a brainstorm session **Already has a specific idea?** Skip this skill. Use the product-agent skill instead. ## What This Skill Does ### 1. Developer Profile Elicitation Gathers context about the developer to personalize ideas: - **Technical skills** — Languages, frameworks, platforms (Swift, SwiftUI, UIKit, CloudKit, etc.) - **Domain interests** — Health, finance, productivity, education, creativity, etc. - **Platform preference** — iOS, macOS, iPadOS, watchOS, visionOS, or multi-platform - **Time availability** — Side project (5-10 hrs/week), full-time, weekend hack - **Constraints** — Budget, backend experience, design skills, existing audience ### 2. Five Brainstorming Lenses Each lens generates 5-8 raw ideas. Lenses are applied in order: 1. **Skills & Interests** — What can you uniquely build given what you know? 2. **Problem-First** — What took 30 seconds today that should take 5? What's annoying? 3. **Technology-First** — Which Apple frameworks have few indie apps? (Foundation Models, ActivityKit, WidgetKit, VisionKit, WeatherKit, MapKit) 4. **Market Gap** — Where are App Store categories underserved, stale, or overpriced? 5. **Trend-Based** — What macro trends create new app opportunities? (AI-native, privacy-first, health/wellness, single-use utilities, Apple Silicon productivity) ### 3. Feasibility Filtering Each raw idea is evaluated against five filters: | Filter | Question | |--------|----------| | Solo Dev Scope | Can one developer ship an MVP in 4-8 weeks? | | Platform API Fit | Does this leverage Apple platform APIs well? | | Monetization Viability | Is there a clear path to revenue? | | Competition Density | How crowded is this space? | | Technical Complexity | Does the developer have the skills to build this? | ### 4. Scoring and Ranking Each surviving idea is scored across five dimensions: | Dimension | WEAK (1-2) | MODERATE (3-5) | STRONG (6-8) | EXCELLENT (9-10) | |-----------|------------|----------------|--------------|------------------| | Solo Dev Scope | 6+ months | 3-6 months | 4-8 weeks | 2-4 weeks | | Platform API Fit | Generic/cross-platform | Some native APIs | Good platform integration | Deep Apple API usage | | Monetization | Unclear model | Ad-supported or low ARPU | Subscription viable | Premium pricing justified | | Competition | Dominated by incumbents | Crowded but fragmented | Few quality options | Blue ocean | | Technical Fit | Major skill gaps | Some learning needed | Good skill match | Perfect skill match | **Overall Score** = weighted average on 1-10 scale (Solo Dev Scope and Technical Fit weighted 1.5x). ### 5. Shortlist Output Final output is a ranked list of 3-5 ideas, each with: - One-liner description - Which lens generated it - Problem statement and target user - Feasibility scores - Monetization model - MVP scope estimate - `next_step` description for feeding the idea into the product-agent skill ## Output Structure ```json { "developer_profile": { "skills": ["Swift", "SwiftUI", "HealthKit", "Core Motion"], "interests": ["fitness", "wearables", "data visualization"], "platform": "iOS + watchOS", "time_availability": "side project (10 hrs/week)", "constraints": ["no backend experience", "solo developer"] }, "brainstorm_lenses_used": [ "skills_and_interests", "problem_first", "technology_first", "market_gap", "trend_based" ], "shortlist": [ { "rank": 1, "idea": "Workout Recovery Timer", "lens": "skills_and_interests", "one_liner": "Apple Watch app that tracks heart rate recovery between sets and suggests optimal rest periods", "platform": "watchOS + iOS companion", "problem_statement": "Gym-goers either rest too long (wasting time) or too short (risking injury). No app uses real-time HR data to personalize rest periods.", "target_user": "Regular gym-goers who wear Apple Watch during workouts", "feasibility": { "solo_dev_scope": "STRONG (6 weeks — HealthKit + WatchKit + simple UI)", "platform_api_fit": "EXCELLENT (HealthKit, WorkoutKit, Live Activities)", "monetization_viability": "STRONG (subscription $3.99/mo — fitness users pay for tools)", "competition_density": "STRONG (few apps focus specifically on HR-based rest timing)", "technical_fit": "EXCELLENT (matches HealthKit + Core Motion skills)" }, "overall_score": 8.4, "monetization_model": "Freemium — free with 3 workouts/week, $3.99/mo unlimited + insights", "competition_notes": "Strong Timer+ and Intervals Pro exist but focus on pre-set timers, not adaptive HR-based rest", "mvp_scope": "Watch app with HR monitoring during rest, vibration alert when ready, basic iOS companion for history", "next_step": "Run the product-agent skill with idea: 'Apple Watch app that tracks heart rate recovery between sets and suggests optimal rest periods based on real-time HR data' for watchOS + iOS" }, { "rank": 2, "idea": "Walking Meetings Tracker", "lens": "problem_first", "one_liner": "Track steps, route, and calories during meetings — share walking meeting summaries with attendees", "platform": "iOS + watchOS", "problem_statement": "Walking meetings are popular for health but there's no way to track the health benefit or share it. Calendar apps don't connect to HealthKit.", "target_user": "Knowledge workers and managers who take walking meetings", "feasibility": { "solo_dev_scope": "STRONG (5 weeks — HealthKit + MapKit + Calendar integration)", "platform_api_fit": "EXCELLENT (HealthKit, MapKit, EventKit, WidgetKit)", "monetization_viability": "MODERATE (niche audience, $2.99 one-time or $1.99/mo)", "competition_density": "EXCELLENT (no dedicated walking meeting app exists)", "technical_fit": "STRONG (HealthKit skills transfer, MapKit is new but manageable)" }, "overall_score": 7.6, "monetization_model": "One-time purchase $4.99 with optional $1.99/mo for team features", "competition_notes": "Pedometer apps exist but none integrate with calendar or frame walking as meetings", "mvp_scope": "Start/stop walking meeting, auto-detect from calendar, log steps + route, share summary", "next_step": "Run the product-agent skill with idea: 'iOS app that tracks steps, route, and calories during walking meetings and shares summaries with attendees' for iOS + watchOS" }, { "rank": 3, "idea": "Gym Equipment Wait Time", "lens": "market_gap", "one_liner": "Crowdsourced gym equipment availability — see which machines are free before you go", "platform": "iOS", "problem_statement": "Gyms are crowded at peak hours. No way to know if the squat rack is free without going. Leads to wasted time and frustration.", "target_user": "Gym members at commercial gyms (Planet Fitness, LA Fitness, Equinox)", "feasibility": { "solo_dev_scope": "MODERATE (8 weeks — needs crowdsource mechanics and gym database)", "platform_api_fit": "MODERATE (MapKit for gym location, but core logic is custom)", "monetization_viability": "STRONG (subscription or gym partnership revenue)", "competition_density": "STRONG (no quality solution exists — GymBook is workout logging, not availability)", "technical_fit": "MODERATE (needs backend for crowdsource data — outside current skills)" }, "overall_score": 6.8, "monetization_model": "Freemium — free for 1 gym, $2.99/mo for multiple gyms + predictions", "competition_notes": "Some gyms have their own capacity apps but no cross-gym equipment-level tracking exists", "mvp_scope": "Single gym support, manual check-in/check-out for equipment, peak time predictions", "next_step": "Run the product-agent skill with idea: 'Crowdsourced gym equipment availability app that shows which machines are free at your gym' for iOS" } ], "ideas_filtered_out": [ { "idea": "AI Personal Trainer", "lens": "technology_first", "reason": "Extremely competitive (Fitbod, Future, Hevy). Would need 6+ months to differentiate. Failed solo_dev_scope filter." }, { "idea": "Sleep Architecture Analyzer", "lens": "trend_based", "reason": "Apple's own sleep tracking in watchOS 10+ covers most of this. Failed competition_density filter — competing with platform owner." } ], "recommendation": "Start with Rank 1 (Workout Recovery Timer). It scores highest, leverages your exact skills (HealthKit + Core Motion), has a clear monetization path, and can ship in 6 weeks. Run the product-agent skill to validate the problem before committing." } ``` ## How to Generate Ideas ### Step 1: Gather Developer Profile If the user hasn't provided context, ask using AskUserQuestion: **Questions to ask:** - What programming languages and frameworks do you know? (Swift, SwiftUI, UIKit, etc.) - What topics or domains interest you? (health, finance, productivity, etc.) - Which Apple platforms do you want to target? - How much time can you dedicate? (side project, full-time, weekend hack) - Any constraints? (no backend, solo dev, no design skills, etc.) If the user provides partial info, infer reasonable defaults and state assumptions. ### Step 2: Apply Skills & Interests Lens Generate 5-8 ideas from the intersection of what the developer knows and cares about. **Process:** 1. List the developer's top 3-5 skills 2. List their top 3-5 interests 3. For each skill-interest pair, ask: "What app could a developer with [skill] build for people interested in [interest]?" 4. Prefer ideas that use 2+ of their skills simultaneously **Example:** Swift + HealthKit + fitness interest → workout recovery timer, exercise form checker, fitness challenge app ### Step 3: Apply Problem-First Lens Find real problems worth solving. **Process:** 1. Ask the developer: "What took 30 seconds today that should take 5?" 2. Think about common daily friction points in their domain 3. Search for complaint threads: `"[domain] app sucks" OR "[domain] app wish" site:reddit.com` 4. Look for manual processes people do with spreadsheets or notes **Example:** "I always forget which supplements I took" → supplement tracking app ### Step 4: Apply Technology-First Lens Identify Apple frameworks with few quality indie apps. **High-opportunity frameworks (2026):** - **Foundation Models** — On-device LLM, few indie apps yet - **ActivityKit / Live Activities** — Sports, cooking, transit, workout tracking - **WidgetKit** — Glanceable info apps (many categories underserved) - **VisionKit** — Document scanning, text recognition, visual lookup - **WeatherKit** — Hyper-local weather for niche use cases - **SensorKit / Core Motion** — Movement, posture, activity detection - **DeviceActivity / Screen Time API** — Digital wellness, parental controls **Process:** 1. Pick 2-3 frameworks matching the developer's platform preference 2. Search App Store for apps using that framework 3. Identify gaps — what should exist but doesn't? ### Step 5: Apply Market Gap Lens Use WebSearch to find underserved App Store categories. **Searches to run:** ``` "underserved App Store categories 2026" "indie app opportunities iOS 2026" "App Store category [domain] top apps" (check staleness) "[category] app reviews complaints" site:reddit.com ``` **Signals of opportunity:** - Top apps haven't been updated in 12+ months - Reviews complain about the same issues across multiple apps - Premium pricing ($9.99+) with mediocre ratings (< 4.0) - Category has few apps but clear demand ### Step 6: Apply Trend Lens Identify macro trends creating new app opportunities. **Current trends (2026):** - **AI-native apps** — Apps that couldn't exist without on-device AI - **Privacy-first** — Apps that explicitly don't collect data (selling point) - **Health & wellness** — Mental health, sleep, nutrition, fitness - **Single-use utilities** — Do one thing exceptionally well - **Apple Silicon productivity** — Pro tools for Mac (M-series GPU, Neural Engine) - **Aging population** — Accessibility, health monitoring, simplified tech - **Creator economy** — Tools for content creators, freelancers **Process:** 1. Pick 2-3 trends relevant to the developer's interests 2. For each trend, generate 2-3 ideas that serve the trend 3. Prefer trend + skill intersections ### Step 7: Filter, Score, and Rank **Filter phase:** 1. Collect all raw ideas from Steps 2-6 (typically 25-40 ideas) 2. Apply the five feasibility filters (see Section 3) 3. Remove any idea that scores WEAK on Solo Dev Scope or Technical Fit 4. Remove any idea that scores WEAK on 3+ dimensions **Score phase:** 1. Score each surviving idea on all five dimensions 2. Calculate overall score (Solo Dev Scope and Technical Fit weighted 1.5x) 3. Rank by overall score descending **Output phase:** 1. Take the top 3-5 ideas 2. Write full entries with all fields from Output Structure 3. Format `next_step` as description for the product-agent skill 4. List 1-3 filtered-out ideas with reasons 5. Write a recommendation highlighting Rank 1 and why ## End-to-End Example **Developer profile:** > Maya, backend engineer (Python, some Swift), interested in nutrition and cooking. Wants to build her first iOS app as a side project (8 hrs/week). No design skills. Comfortable with APIs but no HealthKit experience. **Lens application:** 1. **Skills & Interests →** Recipe nutrition calculator (Swift + API skills + nutrition interest), meal prep planner, ingredient substitution finder 2. **Problem-First →** "I never know if my meals hit my protein target" → macro tracker that works from photos, "grocery lists from recipes are always wrong" → smart grocery list 3. **Technology-First →** Foundation Models → describe leftovers and get recipe suggestions, VisionKit → scan nutrition labels and auto-log 4. **Market Gap →** Searched "nutrition app complaints reddit" → users hate manual food logging, want simpler interfaces, tired of ads in free tier 5. **Trend →** AI-native + health/wellness → on-device food recognition, privacy-first nutrition tracking (no cloud data) **After filtering and scoring:** | Rank | Idea | Score | Key Strength | |------|------|-------|-------------| | 1 | AI Leftover Recipe Finder | 8.1 | Foundation Models + cooking interest, blue ocean | | 2 | Photo Macro Tracker | 7.4 | Problem-first (everyone hates manual logging) | | 3 | Smart Grocery List | 6.9 | Simple scope, clear monetization | **Recommendation:** "Start with AI Leftover Recipe Finder. It uses Foundation Models (few competitors), aligns with your cooking interest, and can ship a basic version in 5 weeks. Run the product-agent skill to validate demand." ## Integration with Other Skills ``` Phase 0: idea-generator (THIS SKILL) ↓ pick 1 idea from shortlist Phase 1: product-agent → validate the problem ↓ Phase 2: market-research / competitive-analysis → size the opportunity ↓ Phase 3: prd-generator → define features ↓ Phase 4: architecture-spec → technical design ↓ Phase 5+: ux-spec, implementation-guide, test-spec, release-spec ``` The `next_step` field in each shortlist entry describes what to feed into the product-agent skill. After the user picks their favorite idea, they can run it directly to kick off Phase 1 validation. ## When NOT to Use This Skill - **Already has a specific idea** — Go directly to the product-agent skill - **Already validated an idea** — Go to `market-research` or `competitive-analysis` - **Wants code generation** — Use the `generators/` skills - **Building to learn only** — Skip feasibility filtering; just pick what's fun - **Hard deadline** — Skip brainstorming; go straight to the product-agent skill with whatever idea you have ## Deliverables At the end of this skill, you should have: - [ ] Developer profile captured (skills, interests, platform, time, constraints) - [ ] All 5 brainstorming lenses applied - [ ] Feasibility filtering applied (5 filters) - [ ] Shortlist of 3-5 ideas with full scoring - [ ] Ideas ranked by overall_score (highest first) - [ ] Each idea has `next_step` for the product-agent skill - [ ] Filtered-out ideas listed with reasons - [ ] Recommendation paragraph highlighting Rank 1 ## Output File Location Save results to: - `idea-shortlist.json` (project root) **Format:** Use the JSON structure from the Output Structure section. **Next action:** Feed Rank 1 into the product-agent skill using the `next_step` description. --- **Generate fast, filter ruthlessly, validate before committing.** The best app idea is the one that matches your skills, solves a real problem, and can ship in weeks — not months.