User Stories: AI Skill for Product Management
User Stories is an AI skill that writes stories from a real persona's need, not restated tasks. Each story stays small, INVEST, and testable, and gets cut when it adds no clarity.
npx skills add Uxcel-Lab/product-skills --skill pm-user-story
Copied
What this skill does
Stops requirements from turning back into a task list and writes stories that carry a real who, what, and why.
- Frames each story from user value: a persona, an action, and a benefit.
- Grounds every story in a research-based persona and a validated need.
- Keeps the story a small unit of value and pushes detail to acceptance criteria.
- Makes stories INVEST and testable so design, engineering, and QA share one definition of done.
- Maps stories across the journey so each requirement traces to a real need.
- Repairs weak stories by returning to who, what, and why.
- Skips the story entirely when the spec already gives full context.
When to use it
Use it when you need an AI assistant to:
- write user stories from a persona or need;
- frame requirements as stories;
- build a story map;
- apply INVEST to a backlog;
- strengthen stories that read like tasks;
- decide whether a story is worth writing at all.
Decisions this skill helps you make
| Decision | Options |
|---|---|
| Formula vs. free-form | The formula for discrete needs; split or go free-form when one line won't hold |
| Write a story at all | Only when it adds a shared who, what, and why the spec lacks |
| Acceptance-criteria depth | Enough to define done now; fuller criteria when feeding a sprint |
| Single story vs. map | Split by user value; map problem to functionality to story to task |
| Communication framing | Lead with the decision and user value, not design language |
How the skill works
1
Establish the need
The skill checks whose need the story serves, whether it's real, and what it feeds.
2
Apply the core
It frames who, what, and why, keeps the story small, and pushes detail to acceptance criteria.
3
Surface trade-offs
It presents formula versus free-form, splitting choices, and criteria depth with a recommendation for each.
