Claude-skill-registry invest-stories

When to Use

When writing, reviewing, or validating user stories. Used by ARCHITECT-AGENT and PRODUCT-OWNER.

Patterns

INVEST Criteria

I - Independent

✅ Can be developed without waiting for other stories
✅ No circular dependencies
❌ "This needs Story X which needs this"

N - Negotiable

✅ HOW is flexible (implementation not prescribed)
✅ WHAT is clear (outcome defined)
❌ "Must use React Query with exact caching config"

V - Valuable

✅ Delivers value to USER or BUSINESS
✅ Value stated explicitly
❌ "Refactor database layer" (no user value)
✅ "User sees data faster because we optimized queries"

E - Estimable

✅ Team can estimate complexity (S/M/L)
✅ No major unknowns blocking estimation
❌ "Integrate with external API" (which API? what ops?)

S - Small

✅ Completable in 1-3 sessions
✅ Can be code reviewed in one sitting
❌ 10+ acceptance criteria, multiple components

T - Testable

✅ ALL acceptance criteria verifiable
✅ Given/When/Then format used
❌ "System should handle errors gracefully"
✅ "Given invalid input, Then error message X displays"

Anti-Patterns

• Technical stories without user value
• Epic disguised as story (too big)
• Vague AC ("properly handles", "works correctly")
• Implementation prescribed in story
• Circular dependencies between stories

Verification Checklist

• Story traces to PRD requirement
• Each INVEST criterion passes
• AC uses Given/When/Then
• No vague words in AC
• Dependencies are one-way only
• Estimated as S, M, or L