@atlashub/smartstack-cli 3.39.0 → 3.41.0

package/templates/skills/business-analyse/steps/step-03a1-setup.md

@@ -1,506 +1,506 @@
----
-name: step-03a1-setup
-description: Module setup, sections walkthrough, questionnaires, and cross-module references
-model: opus
-next_step: steps/step-03a2-analysis.md
----
-
-> **Context files:** `_shared.md` | `_elicitation.md` | `_architecture.md` | `_module-loop.md`
-> **Reference (5-bis):** `references/entity-architecture-decision.md` — Entity scoring, domain coherence, section patterns
-
-# Step 3a1: Setup - Module Initialization
-
-## MANDATORY EXECUTION RULES
-
-- ALWAYS use ULTRATHINK mode
-- This step is EXECUTED ONCE PER MODULE in topological dependency order
-- ALWAYS check workflow state to determine current module
-- ALWAYS reference completed modules when specifying current one
-- **MANDATORY:** Generate and validate an ASCII/SVG mockup for EVERY section (not optional)
-- ALL questions via AskUserQuestion tool
-- ALL communication in `{language}`
-- NEVER skip per-module validation
-- **ID NAMING RULE (MANDATORY, NO EXCEPTION):**
-  All IDs MUST include a module prefix to guarantee application-wide uniqueness.
-  The prefix is derived from the module code initials (2-4 chars):
-    UserManagement → UM | VehicleManagement → VM | PartsInventory → PI
-    RepairManagement → RM | MaintenanceSchedule → MS | DataSync → DS
-    Notifications → NT | Dashboard → DB | Orders → OR | Customers → CU
-
-  Patterns:
-    UC-{PREFIX}-{NNN}        → UC-RM-001, UC-PI-003
-    BR-{CAT}-{PREFIX}-{NNN}  → BR-VAL-RM-001, BR-CALC-PI-002
-    FR-{PREFIX}-{NNN}        → FR-RM-001
-    OBJ-{PREFIX}-{NNN}       → OBJ-RM-001
-    AC-{PREFIX}-{NNN}        → AC-RM-001
-    RISK-{PREFIX}-{NNN}      → RISK-RM-001
-
-  NEVER use bare IDs (UC-001, BR-VAL-001) in multi-module mode.
-- **SCHEMA CONFORMITY RULE:**
-  ALL data MUST fit within the defined feature-schema.json structure.
-  NEVER create custom top-level fields (KPIDefinitions, ChartConfigurations, etc.)
-  Dashboard modules MUST use specification.dashboards[] (it exists in the schema).
-  If truly needed, use specification.extensions: {} (additionalProperties: true).
-
-## MODE DETECTION
-
-**Check your execution context before proceeding:**
-
-**IF you are running as a TEAM AGENT** (you received a module-specifier prompt from `references/agent-module-prompt.md`):
-  → You are in **AUTONOMOUS mode** (Propose & Review)
-  → Do NOT use `AskUserQuestion` — you PROPOSE, the team lead handles user interaction
-  → Replace ALL `AskUserQuestion` calls with **ULTRATHINK autonomous decisions**
-  → Use the decomposition data + completed module summaries to make informed choices
-  → After completing all steps (03a1 → 03a2 → 03b → 03c → 03d): send `PROPOSAL_READY` to team lead
-  → See `references/agent-module-prompt.md` for autonomous decision rules
-  → See `references/team-orchestration.md` § 3 for communication protocol
-
-**IF you are running in the MAIN CONVERSATION** (classic inline mode):
-  → Normal interactive mode — use `AskUserQuestion` as documented below
-  → No `SendMessage` calls needed
-
-**How to detect:** If your prompt contains `PROPOSE & REVIEW` or `team-lead` as recipient, you are a team agent.
-
----
-
-## YOUR TASK
-
-Specify one module's data model and business logic: walk through sections, define entities, business rules, and validate with client before proceeding to UI specification.
-
----
-
-## EXECUTION SEQUENCE
-
-### 1. Determine Current Module
-
-```
-ba-reader.readApplicationContext({feature_id})
-→ Read metadata.workflow: moduleOrder, currentModuleIndex, completedModules
-→ currentModule = moduleOrder[currentModuleIndex]
-→ Display: "═══ Module {currentModuleIndex + 1}/{moduleOrder.length}: {currentModule} ═══"
-```
-
-IF module already specified (status = "specified" in master):
-  Increment currentModuleIndex, re-check
-  IF all modules specified → Load step-04a-collect.md, STOP
-
-### 1b. Cache Warming for Module Loop (FIRST MODULE ONLY)
-
-> **Performance Optimization:** Pre-load module specification references to reduce redundant reads across all modules.
-> This step runs ONLY for the FIRST module (currentModuleIndex = 0), not repeated for subsequent modules.
-
-```
-IF currentModuleIndex === 0:
-  // Pre-load module specification references (Bucket 3)
-  const moduleRefs = [
-    "~/.claude/skills/business-analyse/references/spec-auto-inference.md",
-    "~/.claude/skills/business-analyse/references/ui-resource-cards.md",
-    "~/.claude/skills/business-analyse/references/ui-dashboard-spec.md",
-    "~/.claude/skills/business-analyse/references/cadrage-pre-analysis.md"
-  ];
-
-  for (const file of moduleRefs) {
-    read(file);  // Pre-load into cache
-  }
-
-  Display: "✓ Cache warmed: module specification references (23KB, 4 files)"
-  Display: "  Expected token savings: 75% reduction in reference reads across module loop"
-  Display: "  Retention: until step-04 (after all modules specified)"
-ELSE:
-  // Subsequent modules use cached references (no re-load)
-  Display: "✓ Using cached module references (from first module)"
-```
-
-**Rationale:**
-
-- Module specification references are read 2-3× PER MODULE (5 modules = 10-15 reads without caching)
-- Pre-loading once at start of module loop eliminates 75% of redundant reads
-- Token savings: ~10,000 tokens for a 5-module application
-- Cache retained until step-04 (when consolidation starts, references no longer needed)
-
-See [references/cache-warming-strategy.md](../references/cache-warming-strategy.md) § Bucket 3 for details.
-
-### 2. Initialize Module Feature.json
-
-```
-// Create or load module-level feature.json
-ba-writer.create({
-  scope: "module",
-  applicationRef: {feature_id},
-  moduleCode: {currentModule.code},
-  path: "docs/{app}/{module_code}/business-analyse/v1.0/feature.json"
-})
-
-// Inherit application roles from master
-ba-reader.readApplicationContext({feature_id})
-→ Copy cadrage.applicationRoles → module.applicationContext.applicationRoles
-```
-
-// Update module status in master: "pending" → "in-progress"
-ba-writer.updateModuleStatus({feature_id}, {currentModule.code}, "in-progress")
-
-### 2-ter. Derive Module Discovery Section (MANDATORY)
-
-> **The module feature.json MUST have a `discovery` section.** In multi-module mode, derive it from the parent application's `cadrage`. In single-module mode, it should already exist from step-01.
-
-```
-ba-reader.readApplicationContext({feature_id})
-→ Read cadrage: problem, asIs, toBe, trigger, stakeholders, risks, acceptanceCriteria
-
-// Filter to this module's scope
-moduleScope = cadrage.coverageMatrix.filter(cm => cm.module === currentModule.code)
-moduleStakeholders = cadrage.stakeholders.filter(s => s.tasks relevant to this module)
-moduleRisks = cadrage.risks.filter(r => r related to this module)
-
-ba-writer.enrichSection({
-  featureId: {module_feature_id},
-  section: "discovery",
-  data: {
-    problem: cadrage.problem,
-    asIs: cadrage.asIs,
-    toBe: cadrage.toBe,
-    trigger: cadrage.trigger,
-    stakeholders: moduleStakeholders,
-    scope: {
-      mustHave: moduleScope.filter(s => s.category === "mustHave").map(s => s.item),
-      shouldHave: moduleScope.filter(s => s.category === "shouldHave").map(s => s.item),
-      couldHave: moduleScope.filter(s => s.category === "couldHave").map(s => s.item),
-      outOfScope: []
-    },
-    risks: moduleRisks,
-    acceptanceCriteria: cadrage.acceptanceCriteria.filter(ac => relevant to module),
-    codebaseContext: cadrage.codebaseContext
-  }
-})
-```
-
-> **STRUCTURE CARD: discovery** — Same format as cadrage but module-scoped.
-> Fields: `problem`, `asIs`, `toBe`, `trigger`, `stakeholders[]`, `scope{}`, `risks[]`, `acceptanceCriteria[]`, `codebaseContext`, `openQuestions[]`
-
-### 2-bis. Coverage Verification (MANDATORY)
-
-> **Before specifying any module, verify that the coverageMatrix from cadrage covers this module.**
-
-```
-ba-reader.readApplicationContext({feature_id})
-→ Read cadrage.coverageMatrix[]
-→ Filter entries where module = {currentModule.code}
-→ Group by category: mustHave, shouldHave, couldHave
-```
-
-Display to client:
-```
-Module {currentModule}: Coverage from original brief
-
-| # | Requirement | Category |
-|---|-------------|----------|
-| 1 | {item} | {category} |
-| ... | ... | ... |
-
-Total: {N} requirements mapped to this module
-  - Must-have: {count}
-  - Should-have: {count}
-  - Could-have: {count}
-```
-
-**VALIDATION:** If mustHave count = 0 AND module is in must priority:
-  → WARNING: "Module {currentModule} is must-priority but has no must-have requirements in coverageMatrix. Verify cadrage coverage."
-
-**POST-SPECIFICATION CHECK (at end of module, before advancing):**
-  For EACH mustHave item in coverageMatrix for this module:
-    → Verify at least 1 UC exists that covers this requirement
-    → If uncovered: BLOCK and ask client to either add a UC or reclassify the requirement
-
-### 3. Section Walkthrough with Client
-
-> **This is the key interactive phase aligned with the 4-level SmartStack hierarchy.**
-> Level 2 = Module, Level 3 = Section, Level 4 = Resource
-
-#### 3a. Identify Module Sections
-
-For each module, propose standard sections based on module type:
-
-> **RULE:** Sections = functional zones only. `list` is ALWAYS present (default).
-> `create`/`edit` = separate pages with own URL routes (`/create` and `/:id/edit`). `detail` = tabbed page from list click.
-
-| Module Type | Sections (functional zones) | Detail page tabs (from list click) |
-|-------------|---------------------------|-----------------------------------|
-| data-centric | list | Infos, {relations} |
-| workflow | list, approve | Infos, {relations}, Historique |
-| integration | list | Infos, Config, Logs, Sync |
-| reporting | dashboard | — |
-| full-module | list, dashboard, approve | Infos, {relations}, Historique |
-
-**Section `list` is automatic and ALWAYS included.** Ask ONLY about additional functional sections:
-
-Ask via AskUserQuestion:
-
-```
-question: "{language == 'fr'
-  ? 'Le module {currentModule} aura une page liste (avec création et détail par onglets). Quelles sections fonctionnelles supplémentaires souhaitez-vous ?'
-  : 'Module {currentModule} will have a list page (with create and tabbed detail). Which additional functional sections do you want?'}"
-header: "Sections"
-multiSelect: true
-options:
-  - label: "Dashboard"
-    description: "{language == 'fr' ? 'Tableau de bord avec KPIs, graphiques et métriques clés' : 'Dashboard with KPIs, charts and key metrics'}"
-  - label: "Validation/Approbation"
-    description: "{language == 'fr' ? 'File d\'approbation avec workflow de changement de statut' : 'Approval queue with status change workflow'}"
-  - label: "Import en masse"
-    description: "{language == 'fr' ? 'Import CSV/Excel avec mapping et validation' : 'CSV/Excel import with mapping and validation'}"
-  - label: "{language == 'fr' ? 'Aucune (liste seule)' : 'None (list only)'}"
-    description: "{language == 'fr' ? 'Le module n\'a besoin que de la page liste avec détail' : 'Module only needs the list page with detail'}"
-```
-
-#### 3a-depth. Determine Specification Depth
-
-> **Based on module complexity and type, determine how deeply to specify sections and resources.**
-> **This avoids over-specifying simple modules and under-specifying complex ones.**
-
-Read from application master:
-```
-ba-reader.readApplicationContext({feature_id})
-→ module = modules.find(m => m.code === currentModule)
-→ complexity = module.estimatedComplexity   // simple | medium | complex
-→ featureType = module.featureType          // data-centric | workflow | integration | reporting | full-module
-```
-
-**Depth decision matrix:**
-
-| Complexity | featureType | Depth | Behavior |
-|---|---|---|---|
-| simple | data-centric | **convention** | Auto-generate sections + resources from entities. Quick client validation. |
-| simple | reporting | **convention** | Standard dashboard template. Quick validation. |
-| simple | * | **convention** | Standard sections for featureType. Quick validation. |
-| medium | data-centric | **override** | Auto-generate then ask client which sections to customize. |
-| medium | workflow | **full** | State machine required → always full. |
-| medium | * | **override** | Auto-generate then ask client to customize. |
-| complex | * | **full** | Full specification: state machine, typed columns, conditional actions. |
-| * | workflow | **full** | Always full for workflow modules (state machine is mandatory). |
-
-Display to client:
-```
-"Module {currentModule}: complexité {complexity}, type {featureType} → profondeur: {depth}"
-```
-
-**IF depth = convention:**
-  - Execute 3a-infer (auto-inference) → generates sections + resources
-  - Execute 3b (wireframes) for each section
-  - Skip 3a-bis detailed walkthrough → present auto-generated spec for quick validation
-  - Ask client: "La spécification auto-générée convient-elle, ou souhaitez-vous personnaliser certaines sections ?"
-    - If client says OK → proceed to step 4 (questionnaires)
-    - If client wants changes → switch to **override** for specified sections
-
-**IF depth = override:**
-  - Execute 3a-infer (auto-inference) → generates base sections + resources
-  - Ask client: "Quelles sections souhaitez-vous personnaliser ?"
-  - For selected sections → execute full 3a-bis + 3b + 3b-ter
-  - For other sections → keep auto-generated spec
-
-**IF depth = full:**
-  - Execute standard flow: 3a-bis → 3b → 3b-bis → 3b-ter → 3c → 3d (as currently defined below)
-  - PLUS: 3a-state (state machine definition) for entities with status fields
-
-#### 3a-infer. Auto-Infer Resources from Entities (Convention/Override)
-
-> **For convention and override depths, auto-generate resource details from entity definitions.**
-> **The goal: produce a complete spec without manual input, that the client only validates.**
-
-**Prerequisites:** Entity attributes must be defined (from step 6) OR anticipated from decomposition.
-
-See [references/spec-auto-inference.md](../references/spec-auto-inference.md) for complete inference rules:
-- Entity attribute → SmartTable column mapping (9 type rules)
-- Entity attribute → SmartForm field mapping (8 type rules)
-- Auto-generated sections per featureType (5 types)
-- Component generation rules (list section, create/edit form pages with own routes, detail page with tabs, dashboard section)
-- Status/lifecycle enhancement rules
-
-Write auto-generated sections to `specification.sections[]` via ba-writer.enrichSection()
-
-### 4. Conditional Questionnaires
-
-Based on module type, load questionnaires per-module:
-
-| Condition | Category | Questionnaire |
-|-----------|----------|---------------|
-| Always | 04 | `questionnaire/04-data.md` |
-| Has UI sections | 07 | `questionnaire/07-ui.md` |
-| Has lifecycle | 11 | `questionnaire/11-data-lifecycle.md` |
-| Has migration needs | 12 | `questionnaire/12-migration.md` |
-| References other modules | 13 | `questionnaire/13-cross-module.md` |
-| Documentation needed | 10 | `questionnaire/10-documentation.md` |
-
-Ask via AskUserQuestion, 1-2 batches per category.
-
-#### Store Documentation Requirements
-
-If questionnaire 10 was loaded, persist answers in feature.json:
-
-```
-ba-writer.enrichSection({
-  featureId: {feature_id},
-  section: "documentation",
-  data: {
-    userDocRequired: {Q10.1 answer == "Yes"},
-    techDocRequired: {Q10.2 answer == "Yes"},
-    status: "pending"
-  }
-})
-```
-
-This data will be consumed by `/application` step-08 to auto-generate in-app documentation.
-
-### 5. Cross-Module References
-
-> Reference completed modules to help define current one.
-
-```
-completedModules = ba-reader.getCompletedModulesSummary({feature_id})
-→ Returns max 100 lines summary: entity names, key BRs, permissions
-
-Display: "Modules déjà spécifiés : {completedModules.map(m => m.code).join(', ')}"
-```
-
-For each completed module's entities, ask:
-
-```
-question: "Le module {currentModule} référence-t-il des entités de {completedModule} ?"
-header: "Références"
-multiSelect: true
-options:
-  - label: "{Entity1} (FK)"
-    description: "Référence directe via clé étrangère"
-  - label: "{Entity2} (Lookup)"
-    description: "Table de référence / lookup"
-  - label: "Aucune"
-    description: "Pas de référence à {completedModule}"
-```
-
-#### 5-bis. Unresolved Entity Dependencies (New Module Detection)
-
-> **When an entity reference is detected that does NOT belong to any completed or planned module, apply the SAME intelligent Entity Sourcing Pattern as step-01-cadrage (section 4d-bis).**
-> **Reference:** Load [references/entity-architecture-decision.md](../references/entity-architecture-decision.md) for scoring grid, decision tree, and section patterns.
-
-For each entity referenced by {currentModule} that is NOT covered by any existing/planned module:
-
-##### A. ULTRATHINK: Silent Analysis (NO output to client)
-
-> **Same intelligence as cadrage 4d-bis. Score the entity, assess domain coherence, recommend.**
-
-```
-ULTRATHINK — Entity Architecture Analysis for {entity_name} (discovered during specification of {currentModule}):
-
-1. DOMAIN COHERENCE
-   - Application domain: {app_description}
-   - Entity fit: CORE | SUPPORT | PERIPHERAL
-   - Justification: {1 sentence}
-
-2. ENTITY SCORING (5 criteria, 0-3 each — from scoring grid in reference)
-   - Lifecycle:  {score}/3
-   - Actors:     {score}/3
-   - Relations:  {score}/3
-   - Volume:     {score}/3
-   - Autonomy:   {score}/3
-   - TOTAL:      {total}/15 → {classification}
-
-3. PRE-CLASSIFICATION CHECK (from entity patterns in reference, section 10 of suggestion-catalog)
-   - Match? → {yes: which pattern | no: use scoring}
-
-4. RECOMMENDATION
-   - Option: {in_this_module | new_module | external_system | config_list}
-   - Confidence: HIGH | MEDIUM | LOW
-   - Reasoning: {1-2 sentences}
-
-5. IF recommendation = new_module → PRE-BUILD ARCHITECTURE
-   - Module code: {PascalCase}
-   - Section `list` + detail page tabs
-   - Additional sections (ONLY functional zones)
-   - Impact on dependency graph
-```
-
-##### B. Present Analysis & Recommendation
-
-**BEFORE AskUserQuestion, display as markdown:**
-
-```
-### {language == "fr" ? "Analyse" : "Analysis"} : {entity_plural}
-
-{language == "fr"
-  ? "Pendant la spécification de **{currentModule}**, j'ai détecté une référence aux **{entity_plural}** qui ne font partie d'aucun module existant."
-  : "During specification of **{currentModule}**, I detected a reference to **{entity_plural}** not covered by any existing module."}
-
-{language == "fr" ? "**Mon analyse :**" : "**My analysis:**"}
-- {language == "fr" ? "Cohérence domaine" : "Domain coherence"} : **{CORE|SUPPORT|PERIPHERAL}** — {justification}
-- {language == "fr" ? "Score de complexité" : "Complexity score"} : **{score}/15** ({classification})
-- **{language == "fr" ? "Recommandation" : "Recommendation"}** : **{recommended_option}** — {reasoning}
-```
-
-**IF recommendation = new_module AND confidence >= MEDIUM, ALSO display the architecture proposal** (same format as cadrage 4d-bis Phase 2: Application/Module table + Sections table + Detail tabs table + Impact).
-
-##### C. Ask with Recommended Option First
-
-```
-question: "{language == 'fr'
-  ? 'Comment souhaitez-vous gérer les {entity_plural} ?'
-  : 'How do you want to manage {entity_plural}?'}"
-header: "{entity_name}"
-options:
-  (ORDER: recommended option FIRST with "(Recommandé)"/"(Recommended)", then remaining 3)
-
-  IF recommendation = new_module:
-  - label: "{language == 'fr' ? 'Nouveau module dédié (Recommandé)' : 'New dedicated module (Recommended)'}"
-    description: "{language == 'fr'
-      ? 'Module {ModuleCode} avec page liste, détail par onglets, permissions dédiées'
-      : '{ModuleCode} module with list page, tabbed detail, dedicated permissions'}"
-  - label: "{language == 'fr' ? 'Dans ce module' : 'In this module'}"
-    description: "{language == 'fr'
-      ? 'Onglet dans la page détail de {currentModule}'
-      : 'Tab in {currentModule} detail page'}"
-  - label: "{language == 'fr' ? 'Système externe' : 'External system'}"
-    description: "..."
-  - label: "{language == 'fr' ? 'Liste de référence' : 'Reference list'}"
-    description: "..."
-
-  IF recommendation = config_list:
-  - label: "{language == 'fr' ? 'Liste de référence (Recommandé)' : 'Reference list (Recommended)'}"
-    description: "{language == 'fr'
-      ? 'Table de lookup configurable par l\'admin'
-      : 'Admin-configurable lookup table'}"
-  - (3 remaining options in logical order)
-
-  IF recommendation = in_this_module:
-  - label: "{language == 'fr' ? 'Dans ce module (Recommandé)' : 'In this module (Recommended)'}"
-    description: "{language == 'fr'
-      ? 'Onglet dans la page détail de {currentModule}'
-      : 'Tab in {currentModule} detail page'}"
-  - (3 remaining options)
-
-  IF confidence = LOW:
-  (all 4 options in neutral order, NO "(Recommandé)" tag)
-```
-
-##### D. Process Answer
-
-| Choice | Action |
-|--------|--------|
-| **Dans ce module** | Entity becomes a sub-entity of {currentModule}. Add as a **tab in the detail page**. Add to this module's `analysis.entities[]`. |
-| **Nouveau module dédié** | Add new module to master `modules[]` via ba-writer with proposed architecture (sections, detail tabs, dependencies). The `list` section is ALWAYS the main section. Additional sections ONLY for functional zones. |
-| **Système externe** | Flag for integration. Add to `coverageMatrix` as integration. |
-| **Liste de référence** | Lookup table, admin-managed. No module, no section. |
-
-**IF "Nouveau module dédié" is selected:**
-1. Add new module to master `modules[]` via ba-writer (status: "pending", sortOrder: after current module)
-2. Update dependency graph: current module depends on the new module
-3. **Display:**
-   ```
-   "{language == 'fr'
-     ? '⚠ Nouveau module {ModuleCode} ajouté au graphe de dépendances. Il sera spécifié après {currentModule}.'
-     : '⚠ New module {ModuleCode} added to dependency graph. It will be specified after {currentModule}.'}"
-   ```
-4. Update `metadata.workflow.moduleOrder` to include the new module (inserted after its dependents)
-
----
-
-## NEXT STEP
-
-Load: `./step-03a2-analysis.md`
+---
+name: step-03a1-setup
+description: Module setup, sections walkthrough, questionnaires, and cross-module references
+model: opus
+next_step: steps/step-03a2-analysis.md
+---
+
+> **Context files:** `_shared.md` | `_elicitation.md` | `_architecture.md` | `_module-loop.md`
+> **Reference (5-bis):** `references/entity-architecture-decision.md` — Entity scoring, domain coherence, section patterns
+
+# Step 3a1: Setup - Module Initialization
+
+## MANDATORY EXECUTION RULES
+
+- ALWAYS use ULTRATHINK mode
+- This step is EXECUTED ONCE PER MODULE in topological dependency order
+- ALWAYS check workflow state to determine current module
+- ALWAYS reference completed modules when specifying current one
+- **MANDATORY:** Generate and validate an ASCII/SVG mockup for EVERY section (not optional)
+- ALL questions via AskUserQuestion tool
+- ALL communication in `{language}`
+- NEVER skip per-module validation
+- **ID NAMING RULE (MANDATORY, NO EXCEPTION):**
+  All IDs MUST include a module prefix to guarantee application-wide uniqueness.
+  The prefix is derived from the module code initials (2-4 chars):
+    UserManagement → UM | VehicleManagement → VM | PartsInventory → PI
+    RepairManagement → RM | MaintenanceSchedule → MS | DataSync → DS
+    Notifications → NT | Dashboard → DB | Orders → OR | Customers → CU
+
+  Patterns:
+    UC-{PREFIX}-{NNN}        → UC-RM-001, UC-PI-003
+    BR-{CAT}-{PREFIX}-{NNN}  → BR-VAL-RM-001, BR-CALC-PI-002
+    FR-{PREFIX}-{NNN}        → FR-RM-001
+    OBJ-{PREFIX}-{NNN}       → OBJ-RM-001
+    AC-{PREFIX}-{NNN}        → AC-RM-001
+    RISK-{PREFIX}-{NNN}      → RISK-RM-001
+
+  NEVER use bare IDs (UC-001, BR-VAL-001) in multi-module mode.
+- **SCHEMA CONFORMITY RULE:**
+  ALL data MUST fit within the defined feature-schema.json structure.
+  NEVER create custom top-level fields (KPIDefinitions, ChartConfigurations, etc.)
+  Dashboard modules MUST use specification.dashboards[] (it exists in the schema).
+  If truly needed, use specification.extensions: {} (additionalProperties: true).
+
+## MODE DETECTION
+
+**Check your execution context before proceeding:**
+
+**IF you are running as a TEAM AGENT** (you received a module-specifier prompt from `references/agent-module-prompt.md`):
+  → You are in **AUTONOMOUS mode** (Propose & Review)
+  → Do NOT use `AskUserQuestion` — you PROPOSE, the team lead handles user interaction
+  → Replace ALL `AskUserQuestion` calls with **ULTRATHINK autonomous decisions**
+  → Use the decomposition data + completed module summaries to make informed choices
+  → After completing all steps (03a1 → 03a2 → 03b → 03c → 03d): send `PROPOSAL_READY` to team lead
+  → See `references/agent-module-prompt.md` for autonomous decision rules
+  → See `references/team-orchestration.md` § 3 for communication protocol
+
+**IF you are running in the MAIN CONVERSATION** (classic inline mode):
+  → Normal interactive mode — use `AskUserQuestion` as documented below
+  → No `SendMessage` calls needed
+
+**How to detect:** If your prompt contains `PROPOSE & REVIEW` or `team-lead` as recipient, you are a team agent.
+
+---
+
+## YOUR TASK
+
+Specify one module's data model and business logic: walk through sections, define entities, business rules, and validate with client before proceeding to UI specification.
+
+---
+
+## EXECUTION SEQUENCE
+
+### 1. Determine Current Module
+
+```
+ba-reader.readApplicationContext({feature_id})
+→ Read metadata.workflow: moduleOrder, currentModuleIndex, completedModules
+→ currentModule = moduleOrder[currentModuleIndex]
+→ Display: "═══ Module {currentModuleIndex + 1}/{moduleOrder.length}: {currentModule} ═══"
+```
+
+IF module already specified (status = "specified" in master):
+  Increment currentModuleIndex, re-check
+  IF all modules specified → Load step-04a-collect.md, STOP
+
+### 1b. Cache Warming for Module Loop (FIRST MODULE ONLY)
+
+> **Performance Optimization:** Pre-load module specification references to reduce redundant reads across all modules.
+> This step runs ONLY for the FIRST module (currentModuleIndex = 0), not repeated for subsequent modules.
+
+```
+IF currentModuleIndex === 0:
+  // Pre-load module specification references (Bucket 3)
+  const moduleRefs = [
+    "~/.claude/skills/business-analyse/references/spec-auto-inference.md",
+    "~/.claude/skills/business-analyse/references/ui-resource-cards.md",
+    "~/.claude/skills/business-analyse/references/ui-dashboard-spec.md",
+    "~/.claude/skills/business-analyse/references/cadrage-pre-analysis.md"
+  ];
+
+  for (const file of moduleRefs) {
+    read(file);  // Pre-load into cache
+  }
+
+  Display: "✓ Cache warmed: module specification references (23KB, 4 files)"
+  Display: "  Expected token savings: 75% reduction in reference reads across module loop"
+  Display: "  Retention: until step-04 (after all modules specified)"
+ELSE:
+  // Subsequent modules use cached references (no re-load)
+  Display: "✓ Using cached module references (from first module)"
+```
+
+**Rationale:**
+
+- Module specification references are read 2-3× PER MODULE (5 modules = 10-15 reads without caching)
+- Pre-loading once at start of module loop eliminates 75% of redundant reads
+- Token savings: ~10,000 tokens for a 5-module application
+- Cache retained until step-04 (when consolidation starts, references no longer needed)
+
+See [references/cache-warming-strategy.md](../references/cache-warming-strategy.md) § Bucket 3 for details.
+
+### 2. Initialize Module Feature.json
+
+```
+// Create or load module-level feature.json
+ba-writer.create({
+  scope: "module",
+  applicationRef: {feature_id},
+  moduleCode: {currentModule.code},
+  path: "docs/{app}/{module_code}/business-analyse/v1.0/feature.json"
+})
+
+// Inherit application roles from master
+ba-reader.readApplicationContext({feature_id})
+→ Copy cadrage.applicationRoles → module.applicationContext.applicationRoles
+```
+
+// Update module status in master: "pending" → "in-progress"
+ba-writer.updateModuleStatus({feature_id}, {currentModule.code}, "in-progress")
+
+### 2-ter. Derive Module Discovery Section (MANDATORY)
+
+> **The module feature.json MUST have a `discovery` section.** In multi-module mode, derive it from the parent application's `cadrage`. In single-module mode, it should already exist from step-01.
+
+```
+ba-reader.readApplicationContext({feature_id})
+→ Read cadrage: problem, asIs, toBe, trigger, stakeholders, risks, acceptanceCriteria
+
+// Filter to this module's scope
+moduleScope = cadrage.coverageMatrix.filter(cm => cm.module === currentModule.code)
+moduleStakeholders = cadrage.stakeholders.filter(s => s.tasks relevant to this module)
+moduleRisks = cadrage.risks.filter(r => r related to this module)
+
+ba-writer.enrichSection({
+  featureId: {module_feature_id},
+  section: "discovery",
+  data: {
+    problem: cadrage.problem,
+    asIs: cadrage.asIs,
+    toBe: cadrage.toBe,
+    trigger: cadrage.trigger,
+    stakeholders: moduleStakeholders,
+    scope: {
+      mustHave: moduleScope.filter(s => s.category === "mustHave").map(s => s.item),
+      shouldHave: moduleScope.filter(s => s.category === "shouldHave").map(s => s.item),
+      couldHave: moduleScope.filter(s => s.category === "couldHave").map(s => s.item),
+      outOfScope: []
+    },
+    risks: moduleRisks,
+    acceptanceCriteria: cadrage.acceptanceCriteria.filter(ac => relevant to module),
+    codebaseContext: cadrage.codebaseContext
+  }
+})
+```
+
+> **STRUCTURE CARD: discovery** — Same format as cadrage but module-scoped.
+> Fields: `problem`, `asIs`, `toBe`, `trigger`, `stakeholders[]`, `scope{}`, `risks[]`, `acceptanceCriteria[]`, `codebaseContext`, `openQuestions[]`
+
+### 2-bis. Coverage Verification (MANDATORY)
+
+> **Before specifying any module, verify that the coverageMatrix from cadrage covers this module.**
+
+```
+ba-reader.readApplicationContext({feature_id})
+→ Read cadrage.coverageMatrix[]
+→ Filter entries where module = {currentModule.code}
+→ Group by category: mustHave, shouldHave, couldHave
+```
+
+Display to client:
+```
+Module {currentModule}: Coverage from original brief
+
+| # | Requirement | Category |
+|---|-------------|----------|
+| 1 | {item} | {category} |
+| ... | ... | ... |
+
+Total: {N} requirements mapped to this module
+  - Must-have: {count}
+  - Should-have: {count}
+  - Could-have: {count}
+```
+
+**VALIDATION:** If mustHave count = 0 AND module is in must priority:
+  → WARNING: "Module {currentModule} is must-priority but has no must-have requirements in coverageMatrix. Verify cadrage coverage."
+
+**POST-SPECIFICATION CHECK (at end of module, before advancing):**
+  For EACH mustHave item in coverageMatrix for this module:
+    → Verify at least 1 UC exists that covers this requirement
+    → If uncovered: BLOCK and ask client to either add a UC or reclassify the requirement
+
+### 3. Section Walkthrough with Client
+
+> **This is the key interactive phase aligned with the 4-level SmartStack hierarchy.**
+> Level 2 = Module, Level 3 = Section, Level 4 = Resource
+
+#### 3a. Identify Module Sections
+
+For each module, propose standard sections based on module type:
+
+> **RULE:** Sections = functional zones only. `list` is ALWAYS present (default).
+> `create`/`edit` = separate pages with own URL routes (`/create` and `/:id/edit`). `detail` = tabbed page from list click.
+
+| Module Type | Sections (functional zones) | Detail page tabs (from list click) |
+|-------------|---------------------------|-----------------------------------|
+| data-centric | list | Infos, {relations} |
+| workflow | list, approve | Infos, {relations}, Historique |
+| integration | list | Infos, Config, Logs, Sync |
+| reporting | dashboard | — |
+| full-module | list, dashboard, approve | Infos, {relations}, Historique |
+
+**Section `list` is automatic and ALWAYS included.** Ask ONLY about additional functional sections:
+
+Ask via AskUserQuestion:
+
+```
+question: "{language == 'fr'
+  ? 'Le module {currentModule} aura une page liste (avec création et détail par onglets). Quelles sections fonctionnelles supplémentaires souhaitez-vous ?'
+  : 'Module {currentModule} will have a list page (with create and tabbed detail). Which additional functional sections do you want?'}"
+header: "Sections"
+multiSelect: true
+options:
+  - label: "Dashboard"
+    description: "{language == 'fr' ? 'Tableau de bord avec KPIs, graphiques et métriques clés' : 'Dashboard with KPIs, charts and key metrics'}"
+  - label: "Validation/Approbation"
+    description: "{language == 'fr' ? 'File d\'approbation avec workflow de changement de statut' : 'Approval queue with status change workflow'}"
+  - label: "Import en masse"
+    description: "{language == 'fr' ? 'Import CSV/Excel avec mapping et validation' : 'CSV/Excel import with mapping and validation'}"
+  - label: "{language == 'fr' ? 'Aucune (liste seule)' : 'None (list only)'}"
+    description: "{language == 'fr' ? 'Le module n\'a besoin que de la page liste avec détail' : 'Module only needs the list page with detail'}"
+```
+
+#### 3a-depth. Determine Specification Depth
+
+> **Based on module complexity and type, determine how deeply to specify sections and resources.**
+> **This avoids over-specifying simple modules and under-specifying complex ones.**
+
+Read from application master:
+```
+ba-reader.readApplicationContext({feature_id})
+→ module = modules.find(m => m.code === currentModule)
+→ complexity = module.estimatedComplexity   // simple | medium | complex
+→ featureType = module.featureType          // data-centric | workflow | integration | reporting | full-module
+```
+
+**Depth decision matrix:**
+
+| Complexity | featureType | Depth | Behavior |
+|---|---|---|---|
+| simple | data-centric | **convention** | Auto-generate sections + resources from entities. Quick client validation. |
+| simple | reporting | **convention** | Standard dashboard template. Quick validation. |
+| simple | * | **convention** | Standard sections for featureType. Quick validation. |
+| medium | data-centric | **override** | Auto-generate then ask client which sections to customize. |
+| medium | workflow | **full** | State machine required → always full. |
+| medium | * | **override** | Auto-generate then ask client to customize. |
+| complex | * | **full** | Full specification: state machine, typed columns, conditional actions. |
+| * | workflow | **full** | Always full for workflow modules (state machine is mandatory). |
+
+Display to client:
+```
+"Module {currentModule}: complexité {complexity}, type {featureType} → profondeur: {depth}"
+```
+
+**IF depth = convention:**
+  - Execute 3a-infer (auto-inference) → generates sections + resources
+  - Execute 3b (wireframes) for each section
+  - Skip 3a-bis detailed walkthrough → present auto-generated spec for quick validation
+  - Ask client: "La spécification auto-générée convient-elle, ou souhaitez-vous personnaliser certaines sections ?"
+    - If client says OK → proceed to step 4 (questionnaires)
+    - If client wants changes → switch to **override** for specified sections
+
+**IF depth = override:**
+  - Execute 3a-infer (auto-inference) → generates base sections + resources
+  - Ask client: "Quelles sections souhaitez-vous personnaliser ?"
+  - For selected sections → execute full 3a-bis + 3b + 3b-ter
+  - For other sections → keep auto-generated spec
+
+**IF depth = full:**
+  - Execute standard flow: 3a-bis → 3b → 3b-bis → 3b-ter → 3c → 3d (as currently defined below)
+  - PLUS: 3a-state (state machine definition) for entities with status fields
+
+#### 3a-infer. Auto-Infer Resources from Entities (Convention/Override)
+
+> **For convention and override depths, auto-generate resource details from entity definitions.**
+> **The goal: produce a complete spec without manual input, that the client only validates.**
+
+**Prerequisites:** Entity attributes must be defined (from step 6) OR anticipated from decomposition.
+
+See [references/spec-auto-inference.md](../references/spec-auto-inference.md) for complete inference rules:
+- Entity attribute → SmartTable column mapping (9 type rules)
+- Entity attribute → SmartForm field mapping (8 type rules)
+- Auto-generated sections per featureType (5 types)
+- Component generation rules (list section, create/edit form pages with own routes, detail page with tabs, dashboard section)
+- Status/lifecycle enhancement rules
+
+Write auto-generated sections to `specification.sections[]` via ba-writer.enrichSection()
+
+### 4. Conditional Questionnaires
+
+Based on module type, load questionnaires per-module:
+
+| Condition | Category | Questionnaire |
+|-----------|----------|---------------|
+| Always | 04 | `questionnaire/04-data.md` |
+| Has UI sections | 07 | `questionnaire/07-ui.md` |
+| Has lifecycle | 11 | `questionnaire/11-data-lifecycle.md` |
+| Has migration needs | 12 | `questionnaire/12-migration.md` |
+| References other modules | 13 | `questionnaire/13-cross-module.md` |
+| Documentation needed | 10 | `questionnaire/10-documentation.md` |
+
+Ask via AskUserQuestion, 1-2 batches per category.
+
+#### Store Documentation Requirements
+
+If questionnaire 10 was loaded, persist answers in feature.json:
+
+```
+ba-writer.enrichSection({
+  featureId: {feature_id},
+  section: "documentation",
+  data: {
+    userDocRequired: {Q10.1 answer == "Yes"},
+    techDocRequired: {Q10.2 answer == "Yes"},
+    status: "pending"
+  }
+})
+```
+
+This data will be consumed by `/application` step-08 to auto-generate in-app documentation.
+
+### 5. Cross-Module References
+
+> Reference completed modules to help define current one.
+
+```
+completedModules = ba-reader.getCompletedModulesSummary({feature_id})
+→ Returns max 100 lines summary: entity names, key BRs, permissions
+
+Display: "Modules déjà spécifiés : {completedModules.map(m => m.code).join(', ')}"
+```
+
+For each completed module's entities, ask:
+
+```
+question: "Le module {currentModule} référence-t-il des entités de {completedModule} ?"
+header: "Références"
+multiSelect: true
+options:
+  - label: "{Entity1} (FK)"
+    description: "Référence directe via clé étrangère"
+  - label: "{Entity2} (Lookup)"
+    description: "Table de référence / lookup"
+  - label: "Aucune"
+    description: "Pas de référence à {completedModule}"
+```
+
+#### 5-bis. Unresolved Entity Dependencies (New Module Detection)
+
+> **When an entity reference is detected that does NOT belong to any completed or planned module, apply the SAME intelligent Entity Sourcing Pattern as step-01-cadrage (section 4d-bis).**
+> **Reference:** Load [references/entity-architecture-decision.md](../references/entity-architecture-decision.md) for scoring grid, decision tree, and section patterns.
+
+For each entity referenced by {currentModule} that is NOT covered by any existing/planned module:
+
+##### A. ULTRATHINK: Silent Analysis (NO output to client)
+
+> **Same intelligence as cadrage 4d-bis. Score the entity, assess domain coherence, recommend.**
+
+```
+ULTRATHINK — Entity Architecture Analysis for {entity_name} (discovered during specification of {currentModule}):
+
+1. DOMAIN COHERENCE
+   - Application domain: {app_description}
+   - Entity fit: CORE | SUPPORT | PERIPHERAL
+   - Justification: {1 sentence}
+
+2. ENTITY SCORING (5 criteria, 0-3 each — from scoring grid in reference)
+   - Lifecycle:  {score}/3
+   - Actors:     {score}/3
+   - Relations:  {score}/3
+   - Volume:     {score}/3
+   - Autonomy:   {score}/3
+   - TOTAL:      {total}/15 → {classification}
+
+3. PRE-CLASSIFICATION CHECK (from entity patterns in reference, section 10 of suggestion-catalog)
+   - Match? → {yes: which pattern | no: use scoring}
+
+4. RECOMMENDATION
+   - Option: {in_this_module | new_module | external_system | config_list}
+   - Confidence: HIGH | MEDIUM | LOW
+   - Reasoning: {1-2 sentences}
+
+5. IF recommendation = new_module → PRE-BUILD ARCHITECTURE
+   - Module code: {PascalCase}
+   - Section `list` + detail page tabs
+   - Additional sections (ONLY functional zones)
+   - Impact on dependency graph
+```
+
+##### B. Present Analysis & Recommendation
+
+**BEFORE AskUserQuestion, display as markdown:**
+
+```
+### {language == "fr" ? "Analyse" : "Analysis"} : {entity_plural}
+
+{language == "fr"
+  ? "Pendant la spécification de **{currentModule}**, j'ai détecté une référence aux **{entity_plural}** qui ne font partie d'aucun module existant."
+  : "During specification of **{currentModule}**, I detected a reference to **{entity_plural}** not covered by any existing module."}
+
+{language == "fr" ? "**Mon analyse :**" : "**My analysis:**"}
+- {language == "fr" ? "Cohérence domaine" : "Domain coherence"} : **{CORE|SUPPORT|PERIPHERAL}** — {justification}
+- {language == "fr" ? "Score de complexité" : "Complexity score"} : **{score}/15** ({classification})
+- **{language == "fr" ? "Recommandation" : "Recommendation"}** : **{recommended_option}** — {reasoning}
+```
+
+**IF recommendation = new_module AND confidence >= MEDIUM, ALSO display the architecture proposal** (same format as cadrage 4d-bis Phase 2: Application/Module table + Sections table + Detail tabs table + Impact).
+
+##### C. Ask with Recommended Option First
+
+```
+question: "{language == 'fr'
+  ? 'Comment souhaitez-vous gérer les {entity_plural} ?'
+  : 'How do you want to manage {entity_plural}?'}"
+header: "{entity_name}"
+options:
+  (ORDER: recommended option FIRST with "(Recommandé)"/"(Recommended)", then remaining 3)
+
+  IF recommendation = new_module:
+  - label: "{language == 'fr' ? 'Nouveau module dédié (Recommandé)' : 'New dedicated module (Recommended)'}"
+    description: "{language == 'fr'
+      ? 'Module {ModuleCode} avec page liste, détail par onglets, permissions dédiées'
+      : '{ModuleCode} module with list page, tabbed detail, dedicated permissions'}"
+  - label: "{language == 'fr' ? 'Dans ce module' : 'In this module'}"
+    description: "{language == 'fr'
+      ? 'Onglet dans la page détail de {currentModule}'
+      : 'Tab in {currentModule} detail page'}"
+  - label: "{language == 'fr' ? 'Système externe' : 'External system'}"
+    description: "..."
+  - label: "{language == 'fr' ? 'Liste de référence' : 'Reference list'}"
+    description: "..."
+
+  IF recommendation = config_list:
+  - label: "{language == 'fr' ? 'Liste de référence (Recommandé)' : 'Reference list (Recommended)'}"
+    description: "{language == 'fr'
+      ? 'Table de lookup configurable par l\'admin'
+      : 'Admin-configurable lookup table'}"
+  - (3 remaining options in logical order)
+
+  IF recommendation = in_this_module:
+  - label: "{language == 'fr' ? 'Dans ce module (Recommandé)' : 'In this module (Recommended)'}"
+    description: "{language == 'fr'
+      ? 'Onglet dans la page détail de {currentModule}'
+      : 'Tab in {currentModule} detail page'}"
+  - (3 remaining options)
+
+  IF confidence = LOW:
+  (all 4 options in neutral order, NO "(Recommandé)" tag)
+```
+
+##### D. Process Answer
+
+| Choice | Action |
+|--------|--------|
+| **Dans ce module** | Entity becomes a sub-entity of {currentModule}. Add as a **tab in the detail page**. Add to this module's `analysis.entities[]`. |
+| **Nouveau module dédié** | Add new module to master `modules[]` via ba-writer with proposed architecture (sections, detail tabs, dependencies). The `list` section is ALWAYS the main section. Additional sections ONLY for functional zones. |
+| **Système externe** | Flag for integration. Add to `coverageMatrix` as integration. |
+| **Liste de référence** | Lookup table, admin-managed. No module, no section. |
+
+**IF "Nouveau module dédié" is selected:**
+1. Add new module to master `modules[]` via ba-writer (status: "pending", sortOrder: after current module)
+2. Update dependency graph: current module depends on the new module
+3. **Display:**
+   ```
+   "{language == 'fr'
+     ? '⚠ Nouveau module {ModuleCode} ajouté au graphe de dépendances. Il sera spécifié après {currentModule}.'
+     : '⚠ New module {ModuleCode} added to dependency graph. It will be specified after {currentModule}.'}"
+   ```
+4. Update `metadata.workflow.moduleOrder` to include the new module (inserted after its dependents)
+
+---
+
+## NEXT STEP
+
+Load: `./step-03a2-analysis.md`