← ClaudeAtlas

user-story-mappinglisted

Run user story mapping (Jeff Patton style) — turns a vague goal, brief, or messy backlog into a sliced delivery plan with per-persona user stories that act as the blueprint for the team's test playbook (acceptance criteria + E2E contract). Mines existing context first (README, code, tests, issue tracker via MCP) and reuses the tracker's own taxonomy. Produces a design doc, a story map (Markdown + CSV, plus Mermaid when no tracker is defined), a prioritized backlog (WSJF / RICE / MoSCoW), and optional role-hints for UX/UI + architect. Auto-activates when a sister framework enters its Plan stage — Superpowers between `brainstorming` and `writing-plans`, gstack `/office-hours` / `/autoplan` / `/plan-*-review`, or GSD `/gsd discuss` / `/gsd plan-milestone` / Brief / Roadmap authoring. Use when shaping work (discovery, MVP scoping, "what should we build first", PI planning, organizing a backlog, "stories per persona"). Skip for pure engineering — bug fixes, refactors, code review, deploy debugging.
MartinForReal/storymap-skill · ★ 1 · Code & Development · score 80
Install: claude install-skill MartinForReal/storymap-skill
# User Story Mapping Turns a vague goal, brief, or messy backlog into a **sliced, prioritized delivery plan** with per-persona user stories. The map is the team's **test playbook in skeletal form**: every backbone activity becomes an E2E swimlane, every slice-1 story becomes ≥1 Gherkin scenario (pinned down in Steps 4a/4b). It runs as **one re-entrant loop**. This file is the answer-first spine — the routing-and-contract layer. Read it once at invocation start; load a step's reference only when you reach that step; never pre-read a reference or read one twice. ## The loop Each run discovers what already exists, diffs it against the desired map, and **materializes the delta** in place. **Starting from scratch is the same loop with empty data sources** — the diff is against nothing, so the delta is the whole map. There is no greenfield/brownfield branch; only the data differs, never the steps. ``` ① DISCOVER → ② DIFF → ③ APPLY PREFERENCES → ④ SIMULATE → ⑤ INTERVIEW ──approved?──no──► back to ① │ yes ▼ ⑥ BACKBONE → ⑦ GENERATE/UPDATE → ⑧ DERIVE + HAND OFF ``` **Loop ↔ Step mapping** for the table further down: ①=Step 0, ②=part of Step 0/0.5, ③=Step 0 (apply saved prefs), ④=Step 0.3, ⑤=Step 0.4, ⑥=Step 1, ⑦=Steps 2–4b, ⑧=Steps 5–6. 1. **Discover** (Step 0) — scan data sources cheapest-first; t