@atlashub/smartstack-cli 4.18.0 → 4.20.0

package/templates/skills/business-analyse/SKILL.md

@@ -1,327 +1,208 @@
 ---
 name: business-analyse
-description: Iterative multi-module business analysis with progressive context loading, JSON-first architecture, and interactive HTML deliverables.
-argument-hint: "<feature description>"
+description: "Iterative multi-module business analysis with progressive context loading, JSON-first architecture, and interactive HTML deliverables."
+model: opus
+user_invocable: true
 ---
 
-<objective>
-Execute business analysis workflows. This skill produces feature.json files progressively enriched at each step, published to docs/ for web app rendering.
+# /business-analyse — v2
 
-**Three use cases:**
-1. **New application** — full analysis workflow for a new business need (single or multi-application)
-2. **Update** — delta changes on an existing application or module
-3. **Multi-application project** — when a project spans multiple independent applications (e.g., HR Management + Employee Self-Service)
+## Architecture
 
-The skill auto-detects which use case applies by scanning existing features in `docs/{app}/` and `docs/business-analyse/`.
+**2 phases, 7 steps** — "Macro vers Specification"
 
-**Architecture highlights:**
-- Conditional context loading: each step loads only the shared files it needs (~54% context reduction)
-- 12 granular steps (1-2 responsibilities each, 200-500 lines) for maximum LLM attention quality
-- Self-verification at every step transition, POST-CHECKs on critical outputs
-- Templates and references extracted to separate files, loaded on demand
-</objective>
+### Phase 1 — Macro (Identification + Challenge)
 
-<quick_start>
+| Step | File | Model | Role |
+|------|------|-------|------|
+| 00 | `step-00-init.md` | Sonnet | Detection new/update, scan existing, create index.json structure |
+| 01 | `step-01-cadrage.md` | Opus | Context, stakeholders, scope, risks. ~40 core questions |
+| 02 | `step-02-structure.md` | Opus | Identify App > Module > Section > Resource hierarchy, CHALLENGE each element |
 
-```bash
-# New application analysis
-/business-analyse Complete sales management system
+### Phase 2 — Specification (Deep dive + Client enrichment)
 
-# The skill auto-detects if an existing application matches
-# and proposes "new" or "update" accordingly
-/business-analyse Add export PDF to orders module
-```
+| Step | File | Model | Role |
+|------|------|-------|------|
+| 03 | `step-03-specify.md` | Opus | Deep dive per module: entities, BRs, UCs, permissions, screens. Sequential internal loop |
+| 04 | `step-04-consolidate.md` | Opus | Cross-module validation: data model, permission coherence, E2E flows |
+| 05 | `step-05-deploy.md` | Sonnet | Deploy interactive HTML document |
+| 06 | `step-06-review.md` | Opus | Apply client corrections from ba-review.json exported from HTML |
 
-</quick_start>
+### IA + Client Loop
+```
+Phase 1 (steps 00-02) → Phase 2 (steps 03-05) → Client reviews HTML
+                                                       ↓
+                                          Client comments/enriches
+                                                       ↓
+                                          Export ba-review.json
+                                                       ↓
+                                          Step 06 → re-deploy HTML
+                                                       ↓
+                                          /derive-prd → /ralph-loop
+```
 
-<parameters>
+> **PRD generation:** Handoff data and PRD files are generated by the `/derive-prd` skill, not by step-05.
 
-| Flag | Description |
-|------|-------------|
-| (aucun) | Mode standard — analyse complete ou mise a jour |
-| `-review` | Mode review — lit `ba-review.json`, cree une nouvelle version avec les corrections appliquees |
+## JSON Architecture — Granular files + Index
 
-Step-00 handles detection automatically:
-- **New vs Update**: scans `docs/` for existing applications
-- **Single vs Multi-app**: multi-app detection in step-01, application identity confirmed in step-01b (always)
-- **Single vs Multi-module**: determined during step-02 decomposition
-- **Language**: detected from config or asked once
-- **Review mode**: if `-review` flag, routes directly to step-06
+Each module is split into thematic JSON files with an `index.json` that references them:
 
-</parameters>
+```
+docs/{app}/{module}/business-analyse/v1.0/
+  index.json              ← Index with metadata + file references
+  cadrage.json            ← Context, stakeholders, scope, risks (if app-level)
+  entities.json           ← Entities, attributes, relations
+  rules.json              ← Business rules (BRs)
+  usecases.json           ← Use cases (UCs)
+  permissions.json        ← Permissions, roles, access matrix
+  screens.json            ← Interface specs (sections, resources, mockups)
+  validation.json         ← Validation results, checks
+  handoff.json            ← File mapping, APIs, handoff data
+```
 
-<output_structure>
-**All outputs saved to PROJECT directory:**
+### File hierarchy
 
 ```
 docs/
-    business-analyse/v{X.Y}/
-        feature.json              # PROJECT-level master (only if multi-app: applications, dependencies)
-    {app}/
-        index.json                # BA manifest (auto-updated at handoff, consumed by web app viewer)
-        business-analyse/v{X.Y}/
-            feature.json          # APPLICATION-level master (modules, dependencies, roles)
-            ba-interactive.html   # Client-facing interactive review document (deployed at handoff)
-        {app}/{module}/business-analyse/v{X.Y}/
-            feature.json          # MODULE-level detail (entities, UC, FR, permissions, mockups)
+  business-analyse/v1.0/
+    index.json                    ← Project index (multi-app)
+    cadrage.json                  ← Project-level framing
+  {app}/
+    business-analyse/v1.0/
+      index.json                  ← Application index
+      cadrage.json                ← Application framing
+    {module}/
+      business-analyse/v1.0/
+        index.json                ← Module index
+        entities.json, rules.json, usecases.json, ...
+  ba-interactive.html             ← Single unified HTML
+```
+
+### Partial loading
+
+Each step loads only the files it needs via `index.json`:
+- Step 03 (specify): loads `index.json` + `entities.json` for current module only
+- Step 04 (consolidate): loads all `index.json` files for summary counts
+- Step 05 (deploy): assembles all files into unified FEATURE_DATA for HTML
+
+## Workflow Modes
+
+| Mode | Condition | Behavior |
+|------|-----------|----------|
+| `new` | No existing index.json | Full analysis from step 00 |
+| `update` | Existing index.json found | Delta analysis, skip completed sections |
+| `review` | ba-review.json provided | Apply corrections (step 06 only) |
+
+## Context Loading Rules
+
+Each step loads ONLY what it needs:
+
+| Step | Always loaded | Conditionally loaded |
+|------|---------------|---------------------|
+| 00 | `_shared.md` | — |
+| 01 | `_shared.md`, `step-01-cadrage.md` | questionnaires |
+| 02 | `_shared.md`, `step-02-structure.md` | cadrage.json from step 01 |
+| 03 | `_shared.md`, `step-03-specify.md` | current module's index.json + relevant files |
+| 04 | `_shared.md`, `step-04-consolidate.md` | all module index.json files |
+| 05 | `_shared.md`, `step-05-deploy.md` | all JSON files for HTML assembly |
+| 06 | `_shared.md`, `step-06-review.md` | ba-review.json + current feature files |
+
+## Agents
+
+| Agent | Role | Operations |
+|-------|------|------------|
+| `ba-writer` | Write granular JSON files | create, enrichSection, updateIndex, createVersion, applyReview |
+| `ba-reader` | Read via index.json | findFeature, readSection, getModuleStatus, getSummaryForSkill |
+
+## HTML Output
+
+**Single unified HTML** document (`ba-interactive.html`) with:
+- Multi-app navigation (4 levels: App > Module > Section > Resource)
+- Inline editing (UC, BR, entities, sections/resources)
+- HTML mockups from `screens[]` specs (SmartTable, SmartForm, SmartCard, SmartKanban, SmartDashboard)
+- Wireframe fallback from EMBEDDED_ARTIFACTS
+- Comment system with review panel
+- Export: ba-review.json for step 06 loop
+
+### FEATURE_DATA structure
+
+```javascript
+{
+  metadata: { projectName, version, ... },
+  applications: [{ code, name, icon, route, modules: [...] }],
+  cadrage: { context, scope, stakeholders, risks, criteria },
+  modules: [{ code, applicationCode, name, ... }],
+  moduleSpecs: { [code]: { useCases, businessRules, entities, permissions, screens, sections, resources } },
+  consolidation: { interactions, e2eFlows },
+  comments: []
+}
 ```
-> In standalone mode (single app), the project-level file is NOT created.
 
-**Ralph Loop files (generated by step-05):**
+## Step Execution
+
 ```
-.ralph/
-├── prd-{moduleCode}.json  # Per-module task breakdown for /ralph-loop
-└── progress.txt           # Links to BA documents
+STEP 00 (init)
+  → Load: _shared.md
+  → Execute: step-00-init.md
+  → Output: index.json structure, workflow mode detected
+
+STEP 01 (cadrage)
+  → Load: _shared.md, step-01-cadrage.md, questionnaires
+  → Execute: Interactive Q&A with client (~40 questions)
+  → Output: cadrage.json (context, stakeholders, scope, risks, criteria)
+
+STEP 02 (structure)
+  → Load: _shared.md, step-02-structure.md, cadrage.json
+  → Execute: Identify App > Module > Section > Resource hierarchy
+  → Challenge: Question each element's necessity
+  → Output: modules[], applications[], anticipatedSections per module
+
+STEP 03 (specify) — SEQUENTIAL MODULE LOOP
+  → For each module (in dependency order):
+    → Load: _shared.md, step-03-specify.md, module index.json
+    → Execute: entities, BRs, UCs, permissions, screens
+    → Output: entities.json, rules.json, usecases.json, permissions.json, screens.json
+  → No team mode — sequential ensures 100% context per module
+
+STEP 04 (consolidate)
+  → Load: _shared.md, step-04-consolidate.md, all index.json
+  → Execute: Data model validation, permission coherence, E2E flows
+  → Output: validation.json, consolidated data
+
+STEP 05 (deploy)
+  → Load: _shared.md, step-05-deploy.md, all JSON files
+  → Execute: Assemble FEATURE_DATA, generate HTML
+  → Output: ba-interactive.html (single file)
+
+STEP 06 (review) — CLIENT LOOP ENTRY
+  → Load: _shared.md, step-06-review.md, ba-review.json
+  → Execute: Apply corrections, create new version
+  → Output: Updated JSON files, regenerated HTML
 ```
 
-**Documentation integrated in SmartStack app:**
-- Application view: `/docs/{app}` (module overview, dependency graph, roles)
-- Module view: `/docs/{app}/{module}` (full specification, mockups)
-
-**No intermediate markdown files - all state in feature.json**
-</output_structure>
-
-<navigation_hierarchy>
-**SmartStack uses a 4-level navigation structure mapped to database tables:**
-
-| Level | Entity | Table DB | Created in Step | Description |
-|-------|--------|----------|-----------------|-------------|
-| 1 | Application | `core.nav_Applications` | 00-01 | Top-level application (e.g., "HumanResources") |
-| 2 | Module | `core.nav_Modules` | 02 | Functional module (e.g., "Employees", "TimeManagement") |
-| 3 | **Section** | **`core.nav_Sections`** | **03a-03c** | **Page within module (e.g., "list", "detail", "dashboard")** |
-| 4 | **Resource** | **`core.nav_Resources`** | **03b-03c** | **React component (SmartTable, SmartForm, Chart, KpiCard, etc.)** |
-
-**Hierarchical relationship:**
-- Application → Module → **Section** → **Resource**
-- Each Section MUST have ≥1 Resource
-- Each Section corresponds to 1 React page route
-- Each Resource corresponds to 1 SmartStack UI component
-
-**SeedData generation (step-03c):**
-
-The `specification.seedDataCore` contains **9 mandatory arrays** for database seeding:
-
-| Array | Table | Source | Description |
-|-------|-------|--------|-------------|
-| `navigationApplications` | `core.nav_Applications` | Manual (step-03c) | Application entry (Level 1) — created ONCE per application |
-| `applicationRoles` | `core.auth_Roles` | Manual (step-03c) | Application-scoped roles (admin, manager, contributor, viewer) |
-| `navigationModules` | `core.nav_Modules` | Manual (step-03c) | Module entry (Level 2) |
-| **`navigationSections`** | **`core.nav_Sections`** | **Derived from `specification.sections[]`** | **Section entries (Level 3)** |
-| **`navigationResources`** | **`core.nav_Resources`** | **Derived from `specification.sections[].resources[]`** | **Resource entries (Level 4)** |
-| `navigationTranslations` | `core.nav_Translations` | Manual (i18n) | Multi-language labels |
-| `permissions` | `core.Permissions` | Manual (step-03c) | Permission definitions ({path, action, description}) |
-| `rolePermissions` | `core.RolePermissions` | Manual (step-03c) | Role-permission mappings |
-| `permissionConstants` | `PermissionConstants.cs` | Manual (step-03c) | C# permission constants |
-
-**Transform algorithm (step-03c section 8f-bis):**
-- `navigationSections` = transform `specification.sections[]` → flatten to { code, label, icon, route, parentCode, permission, sort }
-- `navigationResources` = transform `specification.sections[].resources[]` → flatten to { code, type, entity, parentCode, permission }
-
-**Example:**
-- Module: `Employees` (Level 2)
-  - Section: `list` (Level 3) → contains resources:
-    - Resource: `employees-grid` (SmartTable) (Level 4)
-    - Resource: `status-filter` (FilterBar) (Level 4)
-  - Section: `detail` (Level 3) → contains resources:
-    - Resource: `employee-card` (DetailCard) (Level 4)
-    - Resource: `contract-history` (SmartTable) (Level 4)
-
-</navigation_hierarchy>
-
-<resume_workflow>
-**Update mode (auto-detected):**
-
-When step-00 detects that the description matches an existing application:
-
-1. Propose to user: "New application" or "Update {appName}"
-2. If update: locate existing feature.json, create new version (v1.1, etc.)
-3. Load existing feature context
-4. Continue through steps 01-05 with delta focus
-</resume_workflow>
-
-<workflow>
-**Step-00 routing (auto-detection):**
-
-1. Read `{feature_description}`
-2. Scan `docs/` for existing applications
-3. If match found → ask user: "New application" or "Update {app}" ?
-4. If no match → new application
-5. Create master feature.json and proceed to step-01
-
-**New application workflow (single or multi-app):**
-- **Step 00:** Detection, scan existing, create master feature.json (draft)
-- **Step 01:** Cadrage: context, stakeholders, scope, application roles (framed)
-- **Step 01b:** Application identity (always): confirm code, route, icon, seed data (lightweight single-app / full decomposition multi-app)
-- **Step 02:** Decomposition: modules, dependencies, client checkpoint (decomposed) — per application in multi-app mode
-- **Step 03a:** Per-module data: sections, entities, business rules, questionnaires
-- **Step 03b:** Per-module UI: state machines, wireframes, layouts, dashboards
-- **Step 03c:** Per-module compile: actors, UCs, FRs, permissions, navigation, seed data, i18n
-- **Step 03d:** Per-module validate: completeness checks, write feature.json, incremental HTML, loop
-  - Loop: 03a → 03b → 03c → 03d → 03a (next module) until all specified (specified)
-- **Step 04a:** Collect: module summaries, cross-module interactions (FK, events, shared data)
-- **Step 04b:** Analyze: permission coherence, semantic validation, E2E flows, global risks
-- **Step 04c:** Decide: final approval, write consolidation section, proceed to handoff (consolidated)
-- **Step 05a:** Handoff: file mapping (7 categories), BR-to-code mapping, API summary, write to feature.json
-- **Step 05b:** Deploy: prd.json, progress.txt, manifest, ba-interactive.html pre-populated (handed-off)
-- **Step 05c:** Ralph Readiness Check: validation gate before /ralph-loop (optional but recommended)
-
-**Update workflow (same phases, delta focus):**
-- **Step 00:** Detection, locate existing feature, create version N+1
-- **Step 01:** Cadrage delta: what changes, impact on existing
-- **Step 02-05:** Same flow as new, with existing context loaded
-
-**Review workflow (`-review` flag):**
-- **Step 00:** Detection, scan for latest application, locate `ba-review.json`
-- **Step 06:** Apply corrections, create new version, reverse-map data, regenerate all artifacts
-</workflow>
-
-<state_variables>
-**Persist throughout all steps via feature.json:**
-
-| Variable                | Type    | Description                                            |
-| ----------------------- | ------- | ------------------------------------------------------ |
-| `{feature_id}`          | string  | Unique identifier (e.g., `FEAT-001`)                   |
-| `{feature_description}` | string  | What to analyze                                        |
-| `{workflow_type}`       | string  | `"new"` or `"update"`                                  |
-| `{version}`             | string  | SemVer (1.0, 1.1, 2.0, etc.)                           |
-| `{application_name}`    | string  | Target application name                                |
-| `{module_name}`         | string  | Target module name                                     |
-| `{project_id}`          | string  | Project identifier (e.g., `PROJ-001`) — only in multi-app mode |
-| `{workflow_mode}`       | string  | "project" (multi-app), "application" (single-app multi-module), or "module" (single) |
-| `{analysisMode}`        | string  | Always "interactive" — analysis phase is always interactive with the user |
-| `{docs_dir}`            | string  | `docs/{app}/{module}/business-analyse/v{X.Y}/` |
-
-</state_variables>
-
-<entry_point>
+## References
 
-**FIRST ACTION:** Load `steps/step-00-init.md`
+| File | Content |
+|------|---------|
+| `references/html-data-mapping.md` | FEATURE_DATA mapping for HTML generation |
+| `references/naming-conventions.md` | SmartStack naming rules |
+| `references/wireframe-svg-style-guide.md` | SVG wireframe rendering guide |
+| `references/validation-checklist.md` | Validation criteria checklist |
+| `references/deploy-modes.md` | Deployment configuration |
+| `references/acceptance-criteria.md` | Acceptance criteria patterns |
+
+> **Cross-reference:** Handoff data, PRD generation, file templates, and BR-to-code mappings are in the `/derive-prd` skill.
+
+## Questionnaires
 
-</entry_point>
-
-<step_files>
-**Progressive loading - only load current step:**
-
-| Step | File                              | Model  | Purpose                                              |
-| ---- | --------------------------------- | ------ | ---------------------------------------------------- |
-| 00   | `steps/step-00-init.md`           | Sonnet | Auto-detect new/update, scan existing, create master feature.json |
-| 01   | `steps/step-01-cadrage.md`        | Opus   | Framing: context, stakeholders, scope, application roles + multi-app detection |
-| 01b  | `steps/step-01b-applications.md`  | Opus   | Application identity: confirm code, route, icon, seedData (lightweight single-app, full decomposition multi-app) |
-| 02   | `steps/step-02-decomposition.md`  | Opus   | Module decomposition, dependency graph, client checkpoint (per application in multi-app) |
-| 03a1 | `steps/step-03a1-setup.md`        | Opus   | Per-module: setup, sections, questionnaires, cross-refs |
-| 03a2 | `steps/step-03a2-analysis.md`     | Opus   | Per-module: objectives, entities, BRs, process flow |
-| 03b  | `steps/step-03b-ui.md`            | Opus   | Per-module: state machines, wireframes, layouts, dashboards |
-| 03c  | `steps/step-03c-compile.md`       | Opus   | Per-module: actors, UCs, FRs, permissions, navigation, seedDataCore (9 arrays), i18n |
-| 03d  | `steps/step-03d-validate.md`      | Sonnet | Per-module: validation, write, incremental HTML, loop decision |
-| 04a  | `steps/step-04a-collect.md`       | Opus   | Collect module summaries & cross-module interactions |
-| 04b  | `steps/step-04b-analyze.md`       | Opus   | Analyze permission coherence, semantic validation, E2E flows |
-| 04c  | `steps/step-04c-decide.md`        | Opus   | Final approval, write consolidation, proceed to handoff |
-| 05a  | `steps/step-05a-handoff.md`       | Sonnet | Handoff: file mapping (7 categories), BR-to-code mapping, API summary, write handoff |
-| 05b  | `steps/step-05b-deploy.md`        | Sonnet | Deploy: prd.json, progress.txt, manifest, ba-interactive.html |
-| 05c  | `steps/step-05c-ralph-readiness.md` | Opus | ULTRATHINK quality review + ralph readiness gate (MANDATORY) |
-| 06   | `steps/step-06-review.md`         | Opus   | Apply review corrections, create new version, regenerate all artifacts |
-
-</step_files>
-
-<questionnaire_files>
-**Progressive questionnaire loading (step-01 only):**
-
-Load ONLY relevant categories based on feature type:
-
-| Categorie | Fichier                               | Questions | Quand charger |
-| --------- | ------------------------------------- | --------- | ------------- |
-| 00b       | `questionnaire/00b-project.md`        | 5         | Si workflow_mode = "project" (multi-app) |
-| 00        | `questionnaire/00-application.md`     | 18        | Si workflow_mode = "application" ou "project" |
-| 01        | `questionnaire/01-context.md`         | 32        | Toujours      |
-| 02        | `questionnaire/02-stakeholders.md`    | 33        | Toujours      |
-| 03        | `questionnaire/03-scope.md`           | 32        | Toujours      |
-| 04        | `questionnaire/04-data.md`            | 15        | Si module oriente donnees |
-| 05        | `questionnaire/05-integrations.md`    | 14        | Si integrations |
-| 06        | `questionnaire/06-security.md`        | 13        | Si securite mentionnee |
-| 07        | `questionnaire/07-ui.md`              | 19        | Si module oriente interface |
-| 08        | `questionnaire/08-performance.md`     | 8         | Si performance critique |
-| 09        | `questionnaire/09-constraints.md`     | 6         | Si contraintes |
-| 10        | `questionnaire/10-documentation.md`   | 7         | Toujours      |
-| 11        | `questionnaire/11-data-lifecycle.md`  | 14        | Si lifecycle requis (retention, archivage, RGPD) |
-| 12        | `questionnaire/12-migration.md`       | 14        | Si nouveau module (migration donnees existantes) |
-| 13        | `questionnaire/13-cross-module.md`    | 14        | Si nouveau module OU references autres modules |
-| 14        | `questionnaire/14-risk-assumptions.md`| 16        | Toujours      |
-| 15        | `questionnaire/15-success-metrics.md` | 17        | Toujours      |
-
-**Total noyau (toujours chargés) : 137 questions** (contexte, acteurs, périmètre, risques, métriques, documentation)
-**Total avec application : 155 questions** (+ identité application si mode multi-module)
-**Total conditionnel : 117 questions** (data, intégrations, sécurité, UI, performance, contraintes, lifecycle, migration, cross-module)
-**Total maximum possible : 272 questions** (noyau + application + tous conditionnels chargés)
-
-</questionnaire_files>
-
-<template_files>
-**Progressive template loading:**
-
-| Template | File                                | Used in step |
-| -------- | ----------------------------------- | ------------ |
-| schema   | `schemas/feature-schema.json`       | All steps    |
-| spec     | `templates/tpl-frd.md`              | 02           |
-| handoff  | `templates/tpl-handoff.md`          | 05a          |
-| suggestions | `patterns/suggestion-catalog.md` | 01           |
-| interactive | `html/ba-interactive.html`        | 03d (incremental), 05b (final deploy)  |
-| mapping    | `references/html-data-mapping.md`  | 03d, 05b (FEATURE_DATA + EMBEDDED_ARTIFACTS mapping) |
-
-</template_files>
-
-<execution_rules>
-
-- **Load one step at a time** - Only load the current step file
-- **Load shared files conditionally** - Each step loads only the _shared files it needs (see step headers)
-- **Reference files on demand** - Templates and mappings in references/ and templates/ are loaded only when needed
-- **ULTRATHINK** for steps 01, 02, 03a-03c, 04 (critical thinking)
-- **State persisted in feature.json** (NOT markdown files)
-- **Use ba-writer agent** for all JSON writes
-- **Use ba-reader agent** for all JSON reads
-- **Follow next_step directive** at end of each step
-- **Use parallel agents** for codebase exploration
-- **Team mode** (2+ modules): After step-02, load `references/team-orchestration.md` and follow team protocol
-- **Inline mode** (1 module): Classic sequential execution, no team
-- **Project mode** (2+ applications): Application loop wraps module loop — see `_module-loop.md`
-</execution_rules>
-
-<team_mode>
-**Automatic team activation for multi-module applications:**
-
-| Condition | Mode |
-|-----------|------|
-| 1 module | Inline (classic, no team) |
-| 2+ modules | Team mode (Propose & Review) |
-
-**Team mode workflow (Parallel Propose & Review):**
-1. Steps 00-02 execute inline in the main conversation (interactive with user)
-2. After step-02 decomposition: clean up old team data (§1a), then `TeamCreate({ team_name: "ba-{appName}" })` — **CRITICAL: capture the RETURNED team name** (may differ from requested)
-3. For each dependency layer: pre-analyze cross-module touchpoints, then spawn autonomous agents **IN PARALLEL** (all modules in the same layer run simultaneously). Team lead pre-injects coordination notes into each agent's prompt, handles runtime cross-module queries between agents, then presents cross-module interaction analysis + individual module reviews to user.
-4. Between layers: shutdown all layer agents, rebuild completedModulesSummary, spawn next layer
-5. After all modules specified: spawn consolidation agent (steps 04a + 04b), present report for user approval (04c)
-6. After consolidation approved: spawn handoff agent (steps 05a + 05b + 05c)
-7. `TeamDelete()` cleanup
-
-**Why team mode:**
-- Each module agent gets a **fresh context window** (zero accumulation)
-- The team lead stays lightweight (< 30% context even with 10+ modules)
-- Module agents work **autonomously** (no AskUserQuestion — they PROPOSE, user REVIEWS)
-- **Parallel execution**: modules in the same dependency layer run simultaneously (~50-66% wall-clock reduction)
-- **Cross-module coordination**: team lead pre-injects known touchpoints + relays runtime queries between parallel agents
-- Error recovery: if an agent crashes, other agents in the layer continue; crashed module retried independently
-
-**Reference:** `references/team-orchestration.md` (loaded by step-02 after decomposition)
-**Agent prompt template:** `references/agent-module-prompt.md`
-</team_mode>
-
-<success_criteria>
-
-- Master feature.json created and enriched progressively through all steps
-- Module decomposition validated by client (step-02 checkpoint)
-- Each module specified iteratively with client validation
-- ASCII/SVG mockups proposed for each module's sections
-- Cross-module consolidation passed (step-04)
-- Development handoff ready for `/ralph-loop`
-- Ralph Loop prd.json generated (per module or consolidated)
-- All outputs renderable by SmartStack app viewer at `/docs/{app}/`
-</success_criteria>
+| File | Questions | Focus |
+|------|-----------|-------|
+| `questionnaire/01-context.md` | 3 | Business process, friction points, vision |
+| `questionnaire/02-stakeholders-scope.md` | 10 | User profiles, access levels, scope boundaries |
+| `questionnaire/03-data-ui.md` | ~15 | Data entities, UI expectations, dashboards |
+| `questionnaire/04-risks-metrics.md` | — | NOT USED (risks/metrics out of BA scope) |
+| `questionnaire/05-cross-module.md` | ~8 | Cross-module dependencies, integration impact |
+
+## Entry Point
+
+**FIRST ACTION:** Load `steps/step-00-init.md`