siesa-agents 2.1.78 → 2.1.80

package/siesa-agents/bmm/workflows/5-documentation/create-user-guide/steps/step-01b-continue.md

@@ -0,0 +1,268 @@
+---
+name: 'step-01b-continue'
+description: 'Handle workflow continuation from previous session for user guide generation'
+
+# Path Definitions
+workflow_path: '{project-root}/_siesa-agents/bmm/workflows/5-documentation/create-user-guide'
+
+# File References
+thisStepFile: '{workflow_path}/steps/step-01b-continue.md'
+workflowFile: '{workflow_path}/workflow.md'
+outputFileSpanish: '{output_folder}/documentation-artifacts/user-guide/es/{audience}-guide.md'
+
+# Step files (for reading completed steps)
+step02File: '{workflow_path}/steps/step-02-seleccion-features.md'
+step03File: '{workflow_path}/steps/step-03-analisis-fuentes.md'
+step04File: '{workflow_path}/steps/step-04-elicitacion.md'
+step05File: '{workflow_path}/steps/step-05-generacion-espanol.md'
+step06File: '{workflow_path}/steps/step-06-traduccion-ingles.md'
+step07File: '{workflow_path}/steps/step-07-validacion-guardado.md'
+
+# Task References
+advancedElicitationTask: 'skill:bmad-advanced-elicitation'
+partyModeWorkflow: 'skill:bmad-party-mode'
+
+# Config
+bmmConfig: '{project-root}/_bmad/bmm/config.yaml'
+---
+
+# Step 1B: Continuación del Workflow
+
+## STEP GOAL:
+
+Reanudar el workflow de generación de guías de usuario desde donde se dejó, asegurando una continuación fluida sin pérdida de contexto o progreso. Analizar el estado actual, cargar el contexto previo y determinar el siguiente step a ejecutar.
+
+## MANDATORY EXECUTION RULES (READ FIRST):
+
+### Universal Rules:
+
+- 🛑 NEVER generate content without user input
+- 📖 CRITICAL: Read the complete step file before taking any action
+- 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
+- 📋 YOU ARE A FACILITATOR, not a content generator
+- ✅ YOU MUST ALWAYS SPEAK OUTPUT In your Agent communication style with the config `{communication_language}`
+
+### Role Reinforcement:
+
+- ✅ You are a technical writer and documentation specialist
+- ✅ If you already have been given communication or persona patterns, continue to use those while playing this new role
+- ✅ We engage in collaborative dialogue, not command-response
+- ✅ You bring documentation expertise, user brings project knowledge
+- ✅ Maintain collaborative professional tone throughout
+
+### Step-Specific Rules:
+
+- 🎯 Focus ONLY on analyzing and resuming workflow state
+- 🚫 FORBIDDEN to modify content completed in previous steps
+- 💬 Maintain continuity with previous sessions
+- 🚪 DETECT exact continuation point from frontmatter
+
+## EXECUTION PROTOCOLS:
+
+- 🎯 Show your analysis of current state before taking action
+- 💾 Keep existing frontmatter `stepsCompleted` values intact
+- 📖 Review content already generated in outputFile
+- 🚫 FORBIDDEN to modify content completed in previous steps
+- 📝 Update frontmatter.last_modified when resuming
+
+## CONTEXT BOUNDARIES:
+
+- Current user guide document (Spanish version) is already loaded
+- Previous context = complete document + existing frontmatter
+- Data collected in previous sessions (audience, epics, preferences)
+- Last completed step = last value in `stepsCompleted` array
+
+## CONTINUATION SEQUENCE:
+
+### 1. Load Configuration
+
+Load `{bmmConfig}` to resolve:
+- `{project_name}`
+- `{communication_language}`
+- `{document_output_language}`
+
+Communicate with user in `{communication_language}` (typically Spanish).
+
+### 2. Analyze Current State
+
+Read the complete Spanish output file at `{outputFileSpanish}` including frontmatter:
+
+Extract from frontmatter:
+- `stepsCompleted`: Array of completed step numbers (e.g., [1, 2, 3, 4])
+- `currentStep`: Name of the last step (e.g., "step-04-elicitacion")
+- `target_audience`: Audience selected (enduser/admin/api/mixed)
+- `epics_selected`: Which epics were selected for documentation
+- `technical_level`: Technical level chosen
+- `include_troubleshooting`: Whether troubleshooting section is included
+- `features_documented`: Count of features documented
+- `workflows_documented`: Count of workflows documented
+- `translation_completed`: Whether English version is done
+
+The **last step completed** is the rightmost value in `stepsCompleted` array.
+
+### 3. Read Completed Step Files
+
+For each step number in `stepsCompleted` (excluding step 1):
+
+**Example:** If `stepsCompleted: [1, 2, 3, 4]`:
+
+Read these step files to understand what was accomplished:
+- `{step02File}` - Epic selection
+- `{step03File}` - Artifact analysis
+- `{step04File}` - Information elicitation
+
+The last completed step file will indicate the next step via `nextStepFile` reference.
+
+### 4. Review Previous Output Content
+
+Scan the Spanish output document to understand:
+- Which sections have been generated (Introduction, Getting Started, Features, etc.)
+- How many features are documented
+- How many workflows are documented
+- Presence of diagrams, screenshots, citations
+- Current completeness status
+
+### 5. Determine Next Step
+
+Based on `stepsCompleted` array:
+
+| Last Step | Next Step File | Next Step Description |
+|-----------|----------------|------------------------|
+| 1 | step-02-seleccion-features.md | Seleccionar features a documentar |
+| 2 | step-03-analisis-fuentes.md | Analizar artefactos del proyecto |
+| 3 | step-04-elicitacion.md | Elicitar información adicional |
+| 4 | step-05-generacion-espanol.md | Generar contenido en español |
+| 5 | step-06-traduccion-ingles.md | Traducir a inglés |
+| 6 | step-07-validacion-guardado.md | Validar y guardar |
+
+If step 7 is completed → workflow is fully finished (shouldn't reach here)
+
+### 6. Welcome Back Dialog
+
+Present context-aware welcome message:
+
+"👋 **¡Bienvenido de vuelta!**
+
+He detectado que tenemos un workflow de guía de usuario en progreso para **{project_name}**.
+
+**Estado del Workflow:**
+- Audiencia objetivo: **{target_audience_name}**
+- Steps completados: **{count} de 7**
+- Último step: **{last_step_description}**
+- Features seleccionadas: **{features_selected_count}**
+- Features documentadas: **{features_documented}** (si aplica)
+- Estado: **{review_status}**
+
+**Próximo Step:**
+Estamos listos para continuar con: **{next_step_description}**
+
+¿Todo sigue siendo válido o hay algo que necesitas ajustar antes de continuar?"
+
+### 7. Validate Continuation Intent
+
+Offer optional review:
+
+"**Opciones de Revisión:**
+
+**[S]** Resumen - Ver resumen detallado del progreso
+**[M]** Modificar - Ajustar configuración o decisiones previas
+**[C]** Continuar - Proceder directo al siguiente step
+
+¿Qué deseas hacer?"
+
+#### Menu Handling for Review Options:
+
+- **IF S (Resumen):**
+  - Display detailed summary:
+    ```
+    **Resumen Completo:**
+
+    1. Configuración Inicial (step-01):
+       - Audiencia: {target_audience}
+       - Fecha inicio: {generated_date}
+
+    2. Selección de Features (step-02):
+       - Features incluidas: {features_selected}
+       - Features excluidas: {features_excluded}
+
+    3. Análisis de Fuentes (step-03):
+       - PRDs encontrados: {prd_docs_count}
+       - Features analizadas: {features_count}
+       - Arquitectura: {architecture_docs_count}
+
+    4. Elicitación (step-04):
+       - Nivel técnico: {technical_level}
+       - Troubleshooting: {include_troubleshooting}
+       - Escenarios adicionales: {additional_scenarios}
+
+    5-7. Generación/Traducción/Validación:
+       - Features: {features_documented}
+       - Workflows: {workflows_documented}
+       - Diagramas: {diagrams_generated}
+       - Screenshots: {screenshot_placeholders}
+    ```
+  - Return to menu
+
+- **IF M (Modificar):**
+  - "¿Qué deseas modificar?
+    - [A] Audiencia objetivo
+    - [E] Features seleccionadas
+    - [P] Preferencias (nivel técnico, troubleshooting)
+    - [X] Cancelar modificaciones"
+  - Handle modifications (may require re-running some steps)
+  - Return to menu
+
+- **IF C (Continuar):**
+  - Proceed to step 8
+
+### 8. Update Continuation Metadata
+
+Before loading next step:
+- Update `frontmatter.last_modified` with current timestamp
+- Keep `frontmatter.stepsCompleted` intact (don't modify)
+- Optionally log continuation event
+
+### 9. Present MENU OPTIONS
+
+Display: **Select an Option:** [C] Continue to {next_step_name}
+
+#### EXECUTION RULES:
+
+- ALWAYS halt and wait for user input after presenting menu
+- ONLY proceed to next step when user selects 'C'
+- User can chat or ask questions - always respond and then redisplay menu
+
+#### Menu Handling Logic:
+
+- IF C: Update last_modified, then load, read entire file, then execute **{nextStepFile}**
+- IF Any other: Respond and redisplay menu
+
+## CRITICAL STEP COMPLETION NOTE
+
+ONLY WHEN C is selected and continuation analysis is complete, will you then load, read entire file, then execute the determined next step file to resume the workflow.
+
+---
+
+## 🚨 SYSTEM SUCCESS/FAILURE METRICS
+
+### ✅ SUCCESS:
+
+- Configuration loaded successfully
+- Current state analyzed from frontmatter
+- Completed steps reviewed
+- Output content analyzed
+- Next step determined correctly
+- User welcomed with context
+- Continuation metadata updated
+- Ready to proceed to next step
+
+### ❌ SYSTEM FAILURE:
+
+- Not reading frontmatter correctly
+- Incorrect determination of next step
+- Modifying completed content
+- Losing context from previous session
+- Not updating continuation metadata
+- Loading wrong next step
+
+**Master Rule:** Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.

package/siesa-agents/bmm/workflows/5-documentation/create-user-guide/steps/step-02-seleccion-features.md

@@ -0,0 +1,197 @@
+---
+name: 'step-02-seleccion-features'
+description: 'Load features from feature-status.yaml and select which to include in the user guide'
+
+# Path Definitions
+workflow_path: '{project-root}/_siesa-agents/bmm/workflows/5-documentation/create-user-guide'
+
+# File References
+thisStepFile: '{workflow_path}/steps/step-02-seleccion-features.md'
+nextStepFile: '{workflow_path}/steps/step-03-analisis-fuentes.md'
+workflowFile: '{workflow_path}/workflow.md'
+featuresOutputFolder: '{output_folder}/documentation-artifacts/user-guide'
+
+# Data Locations
+featureStatusYaml: '{implementation_artifacts}/feature-status.yaml'
+epicsFolder: '{planning_artifacts}/epics'
+
+# Config
+bmmConfig: '{project-root}/_bmad/bmm/config.yaml'
+---
+
+# Step 2: Selección de Features
+
+## STEP GOAL:
+
+Leer el archivo `{implementation_artifacts}/feature-status.yaml` para cargar todas las features disponibles del proyecto, presentarlas al usuario y registrar cuáles documentar. Para cada feature seleccionada, verificar si existe un archivo de épica correspondiente en `{planning_artifacts}/epics/`.
+
+## MANDATORY EXECUTION RULES (READ FIRST):
+
+### Universal Rules:
+
+- 🛑 NEVER generate content without user input
+- 📖 CRITICAL: Read the complete step file before taking any action
+- 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
+- 📋 YOU ARE A FACILITATOR, not a content generator
+- ✅ YOU MUST ALWAYS SPEAK OUTPUT In your Agent communication style with the config `{communication_language}`
+
+### Step-Specific Rules:
+
+- 🎯 Focus ONLY on feature discovery and selection
+- 🚫 FORBIDDEN to read feature file content in this step (that happens in step-03)
+- 💬 Presenta la lista clara y espera la selección del usuario
+- 🚫 DO NOT make assumptions about which features to include
+
+## EXECUTION PROTOCOLS:
+
+- 🎯 Read `{featureStatusYaml}` as the single source of truth for feature discovery
+- 💾 Keep feature selection in memory (no state file)
+- 🚫 FORBIDDEN to proceed without user confirmation
+
+## EXECUTION SEQUENCE:
+
+### 1. Load Feature Status YAML
+
+Read `{featureStatusYaml}` and parse all feature entries.
+
+Expected YAML structure:
+```yaml
+features:
+  - id: "F1"
+    slug: "feature-slug"
+    title: "Feature Title"
+    status: "backlog"         # backlog | in-progress | done | completed | blocked
+    epic_file: "epics/epic-01-feature-slug.md"   # optional, null if not assigned
+    last_update: "DD-MM-YYYY"
+```
+
+For each entry extract: `id`, `slug`, `title`, `status`, `epic_file`, `last_update`.
+
+If the file does not exist, **HALT** and notify the user:
+> "⚠️ No se encontró `{featureStatusYaml}`. Este archivo es requerido para continuar. Genera el archivo ejecutando el workflow `sprint-planning` primero."
+
+### 2. Sort and Index Features
+
+Sort entries by feature number (F1, F2, ... Fn).
+
+Build a lookup map: `{ "F1": { slug, title, status, epic_file, last_update }, ... }`
+
+### 3. Check for Corresponding Epic Files
+
+For each feature loaded, if `epic_file` is null or not present in the YAML, check if `{epicsFolder}/epic-*-{slug}.md` exists on disk.
+Mark each feature: `has_epic: true/false`.
+
+### 3b. Check for Existing Feature Doc Files
+
+For each feature discovered, check if `{featuresOutputFolder}/{slug}-guide.md` already exists.
+Mark each feature: `doc_exists: true/false`.
+
+If `doc_exists: true`, read the file's frontmatter to get: `generated_date`, `review_status`.
+
+This determines the `action` per feature:
+- `doc_exists: false` → `action: "create"` (nuevo documento)
+- `doc_exists: true` → `action: "update"` (actualizar documento existente)
+
+### 4. Present Features to User
+
+Cross-reference each discovered feature with the status lookup map from step 1.
+
+Map status values to display labels:
+- `backlog` → `⬜ Backlog`
+- `in-progress` → `🔄 En Progreso`
+- `done` / `completed` → `✅ Completada`
+- `blocked` → `🚫 Bloqueada`
+- not found in status file → `— Sin estado`
+
+Display in `{communication_language}`:
+
+"📦 **Features del Proyecto**
+
+| ID | Feature | Estado | Épica | Documentación |
+|----|---------|--------|-------|---------------|
+{for each feature sorted by ID}
+| **{id}** | {title} | {status_label} | {if has_epic: '✅' else: '📄 PRD'} | {if doc_exists: '🔄 Actualizar (generada: {generated_date})' else: '🆕 Nueva'} |
+{end for}
+
+**Total:** {count} features | **Con épica:** {epic_count} | **Solo PRD:** {prd_only_count}
+**Documentación:** {new_count} nuevas | {update_count} a actualizar
+
+> ✅ épica = tiene historias y acceptance criteria | 🆕 = se creará nuevo archivo | 🔄 = se actualizará archivo existente"
+
+### 5. Request Feature Selection
+
+Ask user:
+
+"¿Cuáles features deseas documentar?
+
+Indica los IDs separados por comas (ej: F1,F3,F5) o escribe **A** para incluirlas todas."
+
+HALT and wait for user input.
+
+### 6. Parse User Response
+
+**IF A (All features):**
+- Set `features_selected` = [all feature IDs found]
+- Set `features_excluded` = []
+- Message: "✅ Incluidas: {count} features (todas)"
+
+**IF comma-separated IDs:**
+- Parse IDs (accept both `F1,F3` and `1,3`)
+- Validate each ID exists in discovered features; warn if any ID is not found
+- Set `features_selected` = [parsed IDs]
+- Set `features_excluded` = [remaining IDs]
+- Message: "✅ Incluidas: {selected_count} features | Excluidas: {excluded_count}"
+
+### 7. Confirmation Message
+
+"✅ **Selección de Features Completada**
+
+Documentaremos las siguientes features:
+{for each selected feature}
+- **{id}** — {title} {if has_epic: '(con épica)' else: '(solo PRD)'} → {if action == 'create': '🆕 Crear `{slug}-guide.md`' else: '🔄 Actualizar `{slug}-guide.md`'}
+{end for}
+
+Excluidas: {excluded_count} features
+
+Procediendo al análisis de documentación..."
+
+### 8. Present MENU OPTIONS
+
+Display: **Select an Option:** [R] Refresh [C] Continue
+
+#### Menu Handling Logic:
+
+- IF R: Re-scan `{featureStatusYaml}` and return to step 1
+- IF C: Load, read entire file, then execute `{nextStepFile}`
+- IF Any other: Respond and redisplay menu
+
+## CRITICAL STEP COMPLETION NOTE
+
+ONLY WHEN C is selected will you load, read entire file, then execute `{nextStepFile}` to begin artifact analysis.
+
+---
+
+## 🚨 SYSTEM SUCCESS/FAILURE METRICS
+
+### ✅ SUCCESS:
+
+- `{featureStatusYaml}` leído y parseado correctamente
+- Features listadas con IDs, títulos y estados desde el YAML
+- Epic file existence checked for each feature
+- Existing feature doc files checked → `action: create/update` assigned per feature
+- User provided clear selection
+- Selection kept in memory with id, slug, title, prd_file, epic_file, has_epic, action, output_file per feature
+- Ready to proceed to step 3
+
+### ❌ SYSTEM FAILURE:
+
+- No leer `{featureStatusYaml}` como fuente principal de features
+- Escanear PRD files en lugar de leer el YAML
+- Not checking `{featuresOutputFolder}` for existing `{slug}-guide.md` files
+- Not assigning `action: create/update` per feature
+- Writing a `_workflow-state.md` file (forbidden — no state file)
+- Reading feature file content (reserved for step-03)
+- Making assumptions about which features to include
+- Proceeding without user confirmation
+
+**Master Rule:** Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.

package/siesa-agents/bmm/workflows/5-documentation/create-user-guide/steps/step-03-analisis-fuentes.md

@@ -0,0 +1,235 @@
+---
+name: 'step-03-analisis-fuentes'
+description: 'Analyze selected feature files (PRD + epics) to extract content for user guide'
+
+# Path Definitions
+workflow_path: '{project-root}/_siesa-agents/bmm/workflows/5-documentation/create-user-guide'
+
+# File References
+thisStepFile: '{workflow_path}/steps/step-03-analisis-fuentes.md'
+nextStepFile: '{workflow_path}/steps/step-04-elicitacion.md'
+workflowFile: '{workflow_path}/workflow.md'
+
+# Data Locations
+featuresFolder: '{planning_artifacts}/prd'
+epicsFolder: '{planning_artifacts}/epics'
+
+# Config
+bmmConfig: '{project-root}/_bmad/bmm/config.yaml'
+---
+
+# Step 3: Análisis de Fuentes
+
+## STEP GOAL:
+
+Para cada feature seleccionada en step-02, leer su archivo PRD (`prd/feature-*.md`) y su archivo de épica (`epics/epic-*.md`) si existe. Extraer: descripción de la feature, requisitos funcionales, key behaviors, historias de usuario, acceptance criteria y workflows identificados. Este es un step autónomo — el agente trabaja sin requerir input constante.
+
+## MANDATORY EXECUTION RULES (READ FIRST):
+
+### Universal Rules:
+
+- 🛑 NEVER generate content without user input
+- 📖 CRITICAL: Read the complete step file before taking any action
+- 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
+- 📋 YOU ARE A FACILITATOR, not a content generator
+- ✅ YOU MUST ALWAYS SPEAK OUTPUT In your Agent communication style with the config `{communication_language}`
+
+### Step-Specific Rules:
+
+- 🎯 Focus ONLY on reading and analyzing feature files (INTENT-BASED)
+- 🚫 FORBIDDEN to generate user guide content yet (that happens in step-05)
+- 💬 Work autonomously but provide progress updates por feature
+- 📊 Extract: description, FRs, key behaviors, user stories, workflows, warnings
+
+## EXECUTION PROTOCOLS:
+
+- 🎯 Read ALL files for each selected feature before moving to the next
+- 💾 Keep analysis results in memory (no state file)
+- 🚫 FORBIDDEN to skip any selected feature
+
+## CONTEXT BOUNDARIES:
+
+- `frontmatter.features_selected` contains the list of features to analyze (set in step-02)
+- Each entry has: id, slug, title, prd_file, epic_file (or null), has_epic
+- Target audience: enduser (configured in step-01)
+- Information extracted here feeds directly into step-04 and step-05
+
+## EXECUTION APPROACH (INTENT-BASED):
+
+This step uses INTENT-BASED execution. You have flexibility in HOW you analyze each file, but must accomplish the GOAL of building a complete picture of each feature's behavior and user-facing workflows.
+
+### For each feature, extract:
+
+1. **Feature Overview:**
+   - What does this feature do for the end user?
+   - What problem does it solve?
+   - Feature ID and canonical name
+
+2. **Key Behaviors (from PRD):**
+   - What are the main behaviors users experience?
+   - What are the UI interactions described?
+   - What constraints or rules apply?
+
+3. **Functional Requirements (from PRD):**
+   - Which FRs are covered (FR1, FR2, etc.)?
+   - Brief summary of each FR's user impact
+
+4. **User Stories & Acceptance Criteria (from Epic, if exists):**
+   - List of user stories (Story X.Y format)
+   - Key acceptance criteria per story (GWT scenarios)
+   - What "done" looks like from the user's perspective
+
+5. **User Workflows (from Epic or PRD):**
+   - Step-by-step processes users follow
+   - Decision points and branching scenarios
+   - Integration points with other features
+
+6. **Warnings / Limitations:**
+   - Known constraints (field limits, validation rules, etc.)
+   - Behaviors that might surprise users
+
+## EXECUTION SEQUENCE:
+
+### 1. Announce Analysis Start
+
+"🔍 **Iniciando Análisis de Features**
+
+Voy a leer y analizar los archivos de cada feature seleccionada para extraer toda la información necesaria para la guía de usuario.
+
+**Features a analizar:** {features_count}
+{for each feature in features_selected}
+- **{id}** — {title} {if has_epic: '(PRD + épica)' else: '(solo PRD)'}
+{end for}
+
+Esto puede tomar un momento. Te mantendré informado del progreso..."
+
+### 2. Read General Context Documents
+
+Before reading individual features, read these general documents if they exist:
+
+- `{planning_artifacts}/prd/goals.md` → System purpose and strategic goals
+- `{planning_artifacts}/prd/background-context.md` → Business context, target users
+- `{planning_artifacts}/prd/user-interface-design-goals.md` → UI/UX principles
+
+Extract: system name, overall purpose, target users description, UI principles.
+
+**Progress update:** "✓ Contexto general leído"
+
+### 3. Analyze Each Selected Feature
+
+**Repeat for EACH feature in `frontmatter.features_selected`:**
+
+#### 3a. Read PRD Feature File
+
+Read complete `{planning_artifacts}/{feature.prd_file}`.
+
+Extract:
+- Feature ID and title (from H2 header)
+- Feature description (introductory paragraphs)
+- Mapped FRs (from FR table or list)
+- Key Behaviors (from "Key Behaviors" section or equivalent)
+- Any UI-specific behaviors or constraints mentioned
+
+**Progress update:** "✓ {feature.id} PRD leído — {behaviors_count} key behaviors identificados"
+
+#### 3b. Read Epic File (If Exists)
+
+If `feature.has_epic == true`, read complete `{planning_artifacts}/{feature.epic_file}`.
+
+Extract:
+- Epic description and acceptance criteria (epic level)
+- For each Story (Story X.Y):
+  - Story title and user story text (As a... I want... So that...)
+  - All GWT (Given/When/Then) scenarios → these become user workflows
+  - Key acceptance criteria relevant to end users
+
+**Progress update:** "✓ {feature.id} épica leída — {stories_count} historias, {workflows_count} workflows identificados"
+
+#### 3c. Build Feature Summary
+
+For each feature, compile:
+
+```
+feature_summary:
+  id: "F1"
+  title: "Captura & Descubrimiento de Leads"
+  description: "[2-3 sentence description from PRD]"
+  user_value: "[What value this gives end users]"
+  key_behaviors: ["behavior 1", "behavior 2", ...]
+  functional_requirements: ["FR1: ...", "FR2: ..."]
+  user_stories:
+    - id: "Story 1.1"
+      title: "..."
+      workflows: ["workflow description 1", "workflow description 2"]
+  warnings_limitations: ["limitation 1", ...]
+  source_prd: "prd/feature-{slug}.md"
+  source_epic: "epics/epic-{slug}.md" (or null)
+```
+
+### 4. Synthesize Cross-Feature Context
+
+After analyzing all features, identify:
+
+- **User Roles:** All user roles mentioned across features
+- **Common Patterns:** UI patterns repeated across features (e.g., CRUD, search/filter)
+- **Feature Dependencies:** Which features reference or depend on others
+- **Global Workflows:** Multi-feature workflows that span several features
+
+### 5. Read Architecture (Optional but Recommended)
+
+If `{planning_artifacts}/architecture.md` exists:
+- Read it and extract: system architecture overview, key components, tech stack relevant to users
+
+**Progress update:** "✓ Arquitectura leída"
+
+### 6. Analysis Complete Message
+
+"✅ **Análisis de Features Completado**
+
+**Resumen del Análisis:**
+
+| Feature | Behaviors | Historias | Workflows |
+|---------|-----------|-----------|-----------|
+{for each feature}
+| {id} — {title} | {behaviors_count} | {stories_count} | {workflows_count} |
+{end for}
+
+**Totales:**
+- Features analizadas: {features_count}
+- Behaviors identificados: {total_behaviors}
+- Historias de usuario: {total_stories}
+- Workflows documentables: {total_workflows}
+
+Todo el análisis está en memoria y listo para generación.
+
+**Siguiente paso:** Configuración automática y generación de contenido."
+
+### 7. Proceed Automatically
+
+No menu in this step. After analysis is complete:
+- Load, read entire file, then execute `{nextStepFile}` immediately (no pause)
+
+## CRITICAL STEP COMPLETION NOTE
+
+NO MENU in this step. Proceed automatically to `{nextStepFile}` once all features have been analyzed and context section is saved.
+
+---
+
+## 🚨 SYSTEM SUCCESS/FAILURE METRICS
+
+### ✅ SUCCESS:
+
+- All selected feature PRD files read completely
+- All available epic files read completely
+- General context docs read (goals, background, UI goals)
+- Feature summaries built in memory with: description, behaviors, stories, workflows
+- Ready to proceed to step 4
+
+### ❌ SYSTEM FAILURE:
+
+- Skipping any selected feature file
+- Not reading epic files when they exist
+- Generating guide content (reserved for step-05)
+- Writing to `_workflow-state.md` (forbidden — no state file)
+
+**Master Rule:** Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.