user-stories

user-stories — slice product intent into shippable, testable stories

Turn a PRD/epic/feature into a set of user stories that a team could pick up. Good stories are INVEST: Independent, Negotiable, Valuable, Estimable, Small, Testable. Each delivers user-visible value and has acceptance criteria you could hand to QA.

Process

1. Source the intent

• If a PRD.md exists (e.g. in .plan/<feature-slug>/), read it.
• Otherwise take the epic/feature from the user, and explore the codebase or ask enough to slice responsibly.
• Load a personas skill (e.g. enernet-personas) if available so stories name real personas, not "the user" generically.

2. Build a story map

Lay out the user's journey as a backbone (the big steps), then slice stories under each step. This keeps the set coherent and surfaces gaps. Prefer vertical slices (thin end-to-end value) over horizontal layers (all the backend, then all the frontend).

3. Write each story

Save to <project>/.plan/<feature-slug>/user-stories.md (or where the user prefers). For each:

## US-<n>: <short title>

**As a** <persona>
**I want** <goal/capability>
**So that** <benefit/why>

**Priority:** P0 | P1 | P2
**Depends on:** <other story IDs, or —>

### Acceptance Criteria
- [ ] **Given** <context> **when** <action> **then** <expected outcome>
- [ ] …

### Notes
<edge cases, out-of-scope, open questions for grilling later>

4. Quality bar

• Every story is independently demoable (vertical slice), not "build the schema."
• Acceptance criteria are testable and unambiguous (no "works well").
• Personas are real and specific.
• The set covers the PRD's P0 goals — flag any goal with no story.
• Stories are ordered with dependencies explicit.

5. Hand off

Summarize the story set (count, P0s, ordering). Suggest: "Run /scope to turn these into eng tickets and start the build pipeline."