| name | consumer-product-improvement |
| description | Aggressive consumer-product improvement doctrine. Runs the Analyze → Critique → Reimagine → Stress Test → Improve → Refine cycle on any shipped or in-flight feature. Evaluates UX friction, cognitive load, perceived intelligence, retention mechanics, delight, AI-native opportunities, and five user perspectives (first-time, daily, power, mobile, impatient). Use after every implementation pass, when auditing existing UI, when expanding a feature, when the user shares product feedback, or proactively to elevate "functional but boring" surfaces into premium consumer-grade experiences. Mandates feature-ecosystem thinking, not isolated screens. |
Consumer Product Improvement
When To Use
- After any implementation — never stop at "it works."
- When auditing existing UI — find friction, dead zones, emotional flatness.
- When expanding a feature — think ecosystem, not screen.
- When the user shares product feedback — translate vague reactions into concrete refactors.
- Proactively — the Final Execution Rule (Prompt.md Part II): "Do not merely complete requested tasks. Continuously rethink, improve, optimize, refine, modernize, simplify, and elevate."
The Improvement Cycle
Analyze → Critique → Reimagine → Prototype Mentally → Stress Test → Improve → Refine
| Step | Output |
|---|
| Analyze | What is this surface actually for? What does the user do here? |
| Critique | What's friction, confusion, dead air, generic, slow, missing? |
| Reimagine | What would the best consumer product in the world do here? |
| Prototype Mentally | Walk through the redesigned flow as each of the five personas. |
| Stress Test | Empty data · 10× data · slow network · mobile · keyboard only · accessibility · error states |
| Improve | Ship the highest-ROI changes atomically; one delight per commit. |
| Refine | Polish microinteractions, motion, copy, defaults, transitions. |
The 10 Consumer Questions
For every screen/flow, answer:
- What would make this more useful?
- What would make this feel magical?
- What would reduce friction?
- What would increase delight?
- What would make users return daily?
- What would make users recommend this?
- What would make this feel premium?
- What would save users time or effort?
- What would make this smarter?
- What would users wish existed but never ask for?
Five User Perspectives — Mandatory Walk-through
| Persona | Litmus test |
|---|
| First-time user | Is value obvious in 30 s? Is there a wow moment? Is the next step screaming at them? |
| Daily active user | Is this faster on day 30 than day 1? Does it learn? Does it adapt? |
| Power user | Are shortcuts present? Is there a Cmd+K for everything? Can repetitive work be eliminated? |
| Mobile user | Touch targets ≥ 44 px? No horizontal scroll? Cognitive load minimized? |
| Impatient user | Fewer steps possible? Smart defaults? AI can decide? |
If any persona has friction, the design is not done.
AI-Native Opportunities — Always Scan
- Contextual copilot inline with the current task
- Semantic search across the user's own data
- Predictive recommendations (next action, next experiment, next config)
- Auto-completion on every text field where it adds value
- Proactive insights surfaced unprompted
- Memory across sessions (the product remembers preferences)
- Conversational UX as an alternative to forms
- Personalized dashboards (what this user looks at most → first)
- Intent prediction (the user starts typing → guess the target)
- Auto-summarization of long content (runs, audit logs, eval reports)
- Intelligent prioritization (notifications, lists, dashboards)
AI should be invisible when unnecessary, powerful when needed. Never AI-for-marketing-only.
Delight Engineering Checklist
- Microinteractions on every interactive surface
- Rewarding animations on success (subtle, < 400 ms, GPU-accelerated)
- Smooth transitions between views
- Intelligent empty states (CTA, illustration, sample data offer)
- Playful feedback (success toasts with personality, error states with humanity)
- Premium tactile UX (haptic-style cues, spring physics, easing > linear)
- Frictionless confirmations (optimistic UI, undo > confirm dialogs)
- Smart suggestions (defaults that learn)
- Personalized moments ("Welcome back, your last run …")
- Elegant loading states (skeletons that match the final shape, not spinners)
- Dynamic contextual actions (right-click, long-press, context menus that adapt)
"Would users smile while using this?" — if no, iterate.
Reject
- Generic SaaS dashboards · cluttered UX · feature overload · shallow AI · passive experiences
- "Functional but boring"
- Enterprise-heavy aesthetic in a consumer-facing product
- Anything that feels emotionally flat
Output Format
When this skill is invoked, return:
- Audit summary — top 5–10 opportunities ranked by impact × effort
- Top picks for this iteration — 3–7 improvements to ship now
- Backlog — the rest, sized and tagged
- Verification plan — how each shipped improvement will be validated
Then implement the top picks, commit atomically, verify, push.
