epic-breakdown-advisor

What it does

Breaks down epics into user stories using Humanizing Work patterns in a sequential flow to identify applicable splitting pattern, preserve user value with vertical slices, and evaluate splits for waste.

How it works

1. Step 0: Provide Epic Context by asking for the epic details (title/ID, description, acceptance criteria, target persona, rough estimate).
2. Step 1: Pre-Split Validation (INVEST Check) with prompts to assess independence, negotiability, value, estimability, and testability; proceed only if all checks pass.
3. Step 2: Apply Splitting Patterns Sequentially. For each pattern, determine applicability and generate concrete splits if applicable:
   • Pattern 1: Workflow Steps — deliver thin end-to-end slices rather than step-by-step tasks.
   • Pattern 2: Operations (CRUD) — split into Create, Read, Update, Delete as separate stories when terms imply multiple operations.
   • Pattern 3: Business Rule Variations — create separate stories for different rules/scenarios.
   • Pattern 4: Data Variations — split by data types/structures as variations.
   • Pattern 5: Data Entry Methods — build simplest UI first, add fancy UI later.
   • Pattern 6: Major Effort — first story handles heavy lift; subsequent stories add remaining variants.
   • Pattern 7: Simple/Complex — identify simplest core version first, defer variations.
   • Pattern 8: Defer Performance — deliver functional version first, optimize later.
   • Pattern 9: Break Out a Spike — if none apply, time-box investigation to reduce uncertainty.
4. Step 3: Evaluate Split Quality by asking:
   • Does the split reveal low-value work to deprioritize or eliminate?
   • Do splits produce more equally-sized stories? If not, try a different pattern.
5. Meta-Pattern Application: follow core steps—identify core complexity, list variations, reduce to one complete slice, then separate other variations.

When to use it

Use when a backlog item is too large to estimate, sequence, or deliver safely, to break it into vertical slices that deliver observable user value and reveal potential waste.

What it can touch

The skill interacts via the default interaction protocol for this domain (workshop-facilitation) and asks for epic context, then proposes splits and evaluations; it references patterns and prompts but does not perform external actions itself.

Caveats

No explicit licensing terms are stated beyond NOASSERTION in the metadata. The process requires sequential user input to validate and apply patterns; if a story fails INVEST checks, splitting stops until issues are addressed.