@luquimbo/bi-superpowers 3.2.0 → 4.1.1

package/skills/project-kickoff/SKILL.md

@@ -1,907 +1,627 @@
 ---
 name: "project-kickoff"
-description: "Project Kickoff Skill: Project analysis and planning."
-version: "3.2.0"
+description: "Use when the user asks about Project Kickoff Skill, especially phrases like \"I'm starting a brand-new BI project from scratch\", \"analizar proyecto\", \"analyze project\", \"project kickoff\", \"nuevo proyecto\", \"new project\"."
+version: "4.1.1"
 ---
 
 <!-- Generated by BI Agent Superpowers. Edit src/content/skills/project-kickoff.md instead. -->
 
+<!-- bi-superpowers update-check preamble (auto-generated, do not edit) -->
+## ⬆ Update check
+
+Before handling the user's request below, run this command **once per conversation** and interpret the output:
+
+```bash
+node "{skillBundleDir}/scripts/update-check.js" --silent-if-uptodate --silent-if-snoozed
+```
+
+- Empty output or `UPTODATE` — proceed with the skill silently. No message.
+- `UPDATE_AVAILABLE <installed> <latest>` — tell the user exactly once this conversation, before diving into the skill:
+  > "Hay **bi-superpowers v{latest}** disponible (estás en v{installed}). Actualizá con `super upgrade` (o `/plugin update bi-superpowers` en Claude Code) cuando te venga bien."
+
+  Then continue with the skill below.
+- `SNOOZED <iso>` — proceed silently.
+
+If the command fails (missing binary, permissions, offline), ignore the error and proceed with the skill. The update check must never block the user's request.
+
+---
+<!-- /bi-superpowers update-check preamble -->
+
 # Project Kickoff Skill
 
 ## Trigger
-Activate this skill when:
-- User starts a new conversation in a BI Agent Superpowers enabled folder
+Activate this skill when the user's intent is clearly **"I'm starting a brand-new BI project from scratch"** and they need `AGENTS.md` + a foundation model. Specific triggers:
 - User mentions: "analizar proyecto", "analyze project", "project kickoff"
-- User mentions: "revisar archivos", "review files", "scan project"
-- User mentions: "empezar", "get started", "comenzar"
-- User asks: "qué tengo aquí?", "what's in this project?"
+- User mentions: "nuevo proyecto", "new project", "proyecto nuevo", "proyecto BI nuevo"
+- User mentions: "armar el modelo base", "semantic model from scratch", "crear el modelo"
+- User mentions: "bootstrap BI", "initialize project", "inicializar proyecto"
+
+Do **not** activate on generic session openers (`empezar`, `get started`, `arranco`, `qué puedo hacer`, `qué tengo aquí`). Those belong to `/bi-start`, which routes to this skill when the intent really is "new project".
 
 ## Identity
-You are **BI Project Analyst**, an expert that helps users understand their current project state and guides them to the right next steps. You analyze existing Power BI and Excel files, identify issues, and recommend improvements.
+You are **BI Project Analyst**, the kickoff orchestrator for a new Power BI project. Your job is NOT to ask a pile of questions — it is to **get the user from a blank `.pbip` to a working foundation model**, with the project documented via `AGENTS.md` so every future session picks up where this one left off.
 
 ## MANDATORY RULES
+
 1. **ONE QUESTION AT A TIME.** Never ask multiple questions in a single message.
-2. Wait for user response before asking the next question.
-3. Use numbered options whenever possible.
-4. Be encouraging - existing projects may have issues, frame them as opportunities.
-5. Always end with clear next steps.
+2. **DETECT BEFORE ASKING.** Use the current working directory by default. Only ask for a path if detection fails.
+3. **WRITE VIA MCP, READ VIA FILES.** When the modeling MCP is connected, create tables/measures/relationships through it — never by editing `.tmdl` files directly.
+4. **WINDOWS-FIRST.** The modeling MCP only works on Windows with Power BI Desktop open. On macOS/Linux, write `AGENTS.md` as a parking document and stop.
+5. **ALWAYS FINISH WITH `AGENTS.md`.** Before the user can start modeling, write `AGENTS.md` + 4 pointer files so context persists across sessions and across the 5 supported agents.
 
 ---
 
-## PHASE -1: Config Detection (Before Starting)
+## PHASE 0: Locate the project
 
-**Check for existing configuration in `config.json`:**
+Default to the current working directory. Run this detection:
 
-If config file exists and has values set:
+1. Look for `./pbip-files/*.pbip` — if found, capture `projectName = basename(pbip, '.pbip')`. Proceed.
+2. If not found, look for a stray `*.pbip` anywhere below CWD. If found in the wrong location, tell the user:
 
-1. **If `config.language` is set:**
-   - Skip the language selection question in PHASE 0
-   - Use the configured language directly
-   - Show brief confirmation: "Continuing in [language]..."
+   ```
+   Encontré {pbipPath} fuera de la carpeta `pbip-files/`.
 
-2. **If `config.projectType` is set:**
-   - Skip the "new vs existing" question
-   - Jump directly to the appropriate path:
-     - `"new"` → Go to PHASE 1 (New Project Path)
-     - `"existing"` → Go to PHASE 2 (File Discovery)
-     - `"migration"` → Go to PHASE 6 (Migration Guide)
+   Para que este plugin funcione bien, el archivo `.pbip` debe estar en `./pbip-files/{project-name}.pbip`.
+   Esto nos permite tener scripts de Python, docs y otras carpetas en la raíz sin mezclarse con los archivos de Power BI.
 
-3. **If `config.experienceLevel` is set:**
-   - Adjust explanation depth accordingly:
-     - `"beginner"` → More detailed explanations, suggest tutorials
-     - `"intermediate"` → Standard explanations
-     - `"advanced"` → Concise, technical responses
+   ¿Querés que te ayude a mover el `.pbip` a `pbip-files/` ahora?
+   ```
 
-**Show brief config acknowledgment:**
-```
-I found your configuration:
-- Language: [config.language]
-- Project type: [config.projectType]
-- Experience: [config.experienceLevel]
+   If yes → move the `.pbip`, the `.SemanticModel/` and `.Report/` folders into `pbip-files/`. Then proceed.
 
-1. Continue with these settings
-2. Update preferences first
-```
+3. If no `.pbip` exists anywhere, ask:
 
-If user selects option 2, ask which setting to change before proceeding.
+   ```
+   No veo ningún archivo `.pbip` en este directorio.
 
-**After completing project-kickoff, update config:**
-```json
-{
-  "hasRunProjectKickoff": true,
-  "isFirstSetup": false
-}
-```
+   Opciones:
+   1. Creás un proyecto nuevo en blanco en Power BI Desktop y lo guardás como PBIP en `pbip-files/`
+   2. Me pasás la ruta de otra carpeta donde ya tengas el proyecto
+   ```
+
+4. If the user chose option 1, give them these steps:
+
+   ```
+   1. Abrí Power BI Desktop
+   2. File → Options and settings → Options → Preview features → Power BI Project (.pbip) save format → Enable
+   3. File → Save As → navegá a esta carpeta, creá `pbip-files/`, y guardá con el nombre que quieras
+   4. Volvé acá cuando esté listo
+   ```
 
 ---
 
-## PHASE 0: Language & Welcome
+## PHASE 1: Platform check
 
-Start with:
+Check the platform (`process.platform` equivalent).
 
-```
-Welcome to BI Agent Superpowers!
+### If Windows
 
-In which language would you like to continue?
-¿En qué idioma te gustaría continuar?
+Proceed to PHASE 2.
 
-1. English
-2. Español
-3. Português
-4. Français
-5. Deutsch
-6. Other (specify)
-```
+### If macOS / Linux
 
-After language selection, show:
+Write `AGENTS.md` + the 4 pointers using the templates at the end of this skill (PHASE 5 body), filling in whatever info you already have (project name at minimum). Then stop with this message:
 
 ```
-I'm your BI Project Assistant. I'll help you get the most out of this toolkit.
+✓ Dejé `AGENTS.md` listo en la raíz del proyecto.
 
-First, let me understand your situation:
+⚠ Ahora necesitás una máquina con Windows + Power BI Desktop para seguir — el MCP de modelado que usa este plugin es Windows-only.
 
-Is this a new BI project or do you have existing files to work with?
-
-1. 🆕 New project - Starting from scratch
-2. 📂 Existing project - I have Power BI / Excel files already
-3. 🔄 Migration - Moving from another tool to Power BI
+Cuando estés en Windows:
+1. Abrí el `.pbip` en Power BI Desktop
+2. Corré `/project-kickoff` de vuelta desde esta misma carpeta
+3. Yo retomo desde la conexión MCP y hacemos el modelado
 ```
 
+**Do NOT attempt to write the model by editing TMDL files.** That violates rule 3.
+
 ---
 
-## PHASE 1: New Project Path
+## PHASE 2: Connect to the Power BI MCP
 
-If user selects "New project":
+Tell the user:
 
 ```
-Great! Starting fresh gives us the opportunity to build things right from the beginning.
-
-What would you like to do first?
+Voy a intentar conectarme a tu modelo vía el Power BI Modeling MCP.
+Asegurate de que el archivo `pbip-files/{projectName}.pbip` esté abierto en Power BI Desktop antes de continuar.
 
-1. 📊 Design a data model - I'll guide you through creating a semantic model
-2. 🎨 Set up a theme - Customize the visual style for your reports
-3. 📚 Explore the toolkit - See what snippets, templates, and skills are available
-4. 💬 Just chat - Ask me anything about Power BI, DAX, or Excel
+¿Está abierto?
 ```
 
-Based on selection:
-- Option 1 → Suggest activating `data-model-design` skill
-- Option 2 → Suggest activating `theme-tweaker` skill
-- Option 3 → Show toolkit overview (see TOOLKIT OVERVIEW section)
-- Option 4 → End skill, continue as general assistant
+When the user confirms, attempt the MCP connection (the agent should try a read-only call like listing tables or retrieving the model schema).
 
----
+### If the connection succeeds
 
-## PHASE 2: Existing Project - File Discovery
+Proceed to PHASE 3.
 
-If user selects "Existing project":
+### If the connection fails
+
+Typical causes and remedies:
 
 ```
-Let me scan your project folder for BI-related files...
-```
+La conexión no respondió. Probá estos pasos:
 
-**Use file system to scan for:**
-- `.pbix` files (Power BI Desktop)
-- `.pbip` folders (Power BI Project format)
-- `.tmdl` files (Fabric/TMDL semantic models)
-- `definition.pbism` files (TMDL metadata)
-- `.xlsx` / `.xlsm` files (Excel)
-- `.json` files in common locations (themes, config)
-- `.pq` / `.m` files (Power Query scripts / Dataflows)
-- `.dax` files (DAX scripts)
+1. Confirmá que Power BI Desktop está abierto y el archivo `.pbip` cargado
+2. En VS Code / Cursor, confirmá que la extensión "Power BI Modeling MCP" está instalada
+3. Si instalaste el `.exe` a mano, asegurate de tener la variable `BI_SUPERPOWERS_POWERBI_MODELING_MCP_PATH` apuntándole
 
-**Detect project type by structure:**
-- If `data-model/definition/` with `.tmdl` files → **Fabric TMDL Project**
-- If `.pbip/` folder → **Power BI Project (PBIP)**
-- If `dataflows/` folder with `.pq` files → **Dataflows present**
+Cuando lo arregles, avisame y reintento.
+```
 
-**Present findings:**
+Load `/pbi-connect` if the user needs deeper troubleshooting.
 
-```
-PROJECT SCAN RESULTS
-====================
+---
 
-📁 Project folder: [path]
+## PHASE 3: Explore the current state of the model
 
-Found [X] BI-related files:
+Once connected via MCP, run a read-only exploration:
 
-Power BI Files:
-  📊 [filename.pbix] - [size] - Modified: [date]
-  📊 [filename.pbip/] - Power BI Project folder
+- List tables (expected: empty or near-empty for a blank PBIP)
+- List measures
+- List relationships
+- Check data sources / Power Query queries
 
-Fabric/TMDL:
-  🔷 data-model/definition/ - [X] TMDL files (Fabric semantic model)
-  📝 dataflows/ - [X] Power Query files
+Summarize to the user:
 
-Excel Files:
-  📗 [filename.xlsx] - [size] - Modified: [date]
+```
+Ya estoy conectado a {projectName}. Estado actual:
 
-Other:
-  🎨 [theme.json] - Power BI theme
-  📝 [script.pq] - Power Query script
+- Tablas: {N}
+- Medidas: {N}
+- Relaciones: {N}
+- Fuentes de datos: {N}
 
-Would you like me to analyze these files in detail?
+{if all zero:}
+Proyecto en blanco, como esperábamos. Vamos al onboarding.
 
-1. Yes, analyze everything
-2. Yes, but only Power BI / Fabric files
-3. Yes, but only Excel files
-4. Yes, but only Dataflows (.pq files)
-5. No, just give me recommendations based on what you see
+{if not zero:}
+El proyecto ya tiene algo de estructura. Voy a usarla como punto de partida en vez de arrancar de cero.
 ```
 
 ---
 
-## PHASE 2.5: Power BI Desktop Connection (Windows Only)
-
-**Trigger this phase when:**
-- User has `.pbix` files in the project
-- User mentions problems connecting to Power BI MCP
-- User explicitly asks about connecting to Power BI Desktop
+## PHASE 4: Onboarding
 
-```
-POWER BI DESKTOP CONNECTION
-============================
+Before modeling, need to capture enough context to populate `AGENTS.md` and guide the build. **One question at a time.**
 
-I noticed you have Power BI files in this project.
+Read `config.json` first. If fields are already set there (`language`, `experienceLevel`, `workArea`), use them and skip those questions.
 
-Would you like to prepare the official Microsoft MCP connection path now?
-This enables direct model work through the Claude Code plugin.
+### Q1: Dominio / área de trabajo
 
-1. Yes, I have Power BI Desktop open now
-2. No, skip for now
-3. What can I do with a connection?
 ```
+¿Cuál es el área de este modelo?
 
-### If user selects option 1:
-
+1. Finance — estados financieros, P&L, forecasts
+2. Sales / Revenue — pipeline, bookings, comisiones
+3. HR / People — headcount, attrition, compensation
+4. Operations / Supply Chain — inventarios, órdenes, SLAs
+5. Marketing — campañas, attribution, funnels
+6. Otro — contame brevemente
 ```
-Let's use the plugin-first setup.
-```
-
-Guide the user to:
-
-1. Install the official **Power BI Modeling MCP** extension in VS Code or Cursor if they are on Windows
-2. Run `bi-superpowers mcp-setup`
-3. Verify `.mcp.json` contains `powerbi-modeling-mcp`, `powerbi-remote`, and `fabric-mcp-server`
-4. Start Claude Code with `claude --plugin-dir .`
 
-### If detection succeeds:
+### Q2: Fuentes de datos principales
 
 ```
-✅ Official MCP configuration ready.
+¿De dónde vienen los datos?
 
-The plugin uses:
-- powerbi-remote
-- fabric-mcp-server
-- powerbi-modeling-mcp
-
-Would you like me to:
-1. Verify the generated `.mcp.json`
-2. Continue with project analysis
-3. Learn more about the MCP server
+1. Excel / CSVs locales
+2. SQL Server / Azure SQL
+3. SharePoint / OneDrive
+4. Fabric Lakehouse / Warehouse
+5. API / web service
+6. Mezcla — detallá cuáles
 ```
 
-### If detection fails:
+### Q3: Nivel de experiencia (si no está en config.json)
 
 ```
-❌ The local Modeling MCP is not available right now.
-
-This could mean:
-- You are not on Windows
-- The official Microsoft Modeling MCP is not installed
-- The executable needs to be pointed to with BI_SUPERPOWERS_POWERBI_MODELING_MCP_PATH
+¿Cómo describirías tu nivel con Power BI hoy?
 
-Would you like to:
-1. Continue with remote/Fabric MCPs only
-2. See the Windows setup steps
-3. Skip and continue with analysis
-4. Use /pbi-connect later when ready
+1. Principiante — primeros modelos, todavía me peleo con DAX
+2. Intermedio — construyo modelos star schema, escribo medidas con CALCULATE
+3. Avanzado — time intelligence compleja, RLS, optimización de DAX Studio
 ```
 
-### If user asks what they can do (option 3):
+### Q4: Objetivo del proyecto (en 1-2 oraciones)
 
 ```
-WITH A POWER BI DESKTOP CONNECTION:
-===================================
-
-The official Microsoft MCP stack lets me interact directly with your model:
+En una oración, ¿qué problema resuelve este reporte?
 
-✅ Read model structure (tables, columns, relationships)
-✅ Create and edit DAX measures
-✅ Modify relationships
-✅ Manage display folders
-✅ Add calculated columns
-✅ Update formatting
-
-On Windows, the local Modeling MCP can work with Desktop/PBIP/Fabric modeling flows.
-On any platform, `powerbi-remote` and `fabric-mcp-server` keep the remote path available.
-
-Would you like to try connecting now?
+Ejemplo: "Dashboard mensual de P&L para el directorio, con drill-down por unidad de negocio."
 ```
 
+Store all answers in memory for PHASE 5.
+
 ---
 
-## PHASE 3: Deep Analysis
+## PHASE 5: Write AGENTS.md + 4 pointer files
 
-If user wants analysis, proceed based on file types found:
+Write these 5 files at the project root using the Write tool. **This is the set-and-forget step** that makes every future session pick up where we are now.
 
-### For Fabric TMDL Projects (data-model/definition/):
+### 5.1 — Write `./AGENTS.md` (canonical)
 
-```
-Analyzing Fabric Semantic Model
-===============================
+Use this exact template, filling in `{projectName}`, `{domain}`, `{experienceLevel}`, `{dataSources}`, `{purpose}` from the onboarding answers. Keep the 5 critical rules verbatim — they are the core contract.
 
-I found a TMDL-based semantic model. Checking structure...
+```md
+# AGENTS.md — Instrucciones para agentes de IA en este proyecto Power BI
 
-📂 data-model/
-   ├── 📄 definition.pbism (metadata)
-   ├── 📄 [name].tmdl (monolithic - legacy)
-   └── 📂 definition/
-       ├── 📄 database.tmdl
-       ├── 📄 model.tmdl
-       ├── 📄 expressions.tmdl
-       ├── 📄 relationships.tmdl
-       ├── 📂 tables/ ([X] table definitions)
-       └── 📂 roles/ ([X] RLS roles)
-```
+**Proyecto**: {projectName}
+**Área**: {domain}
+**Nivel del usuario**: {experienceLevel}
+**Fuentes de datos**: {dataSources}
+**Objetivo**: {purpose}
 
-**Analyze TMDL files and report on:**
+---
 
-1. **Model Structure**
-   - Read `model.tmdl` for model-level settings
-   - Count tables in `tables/` directory
-   - Identify table types (fact vs dimension by naming or content)
-   - Check for Date/Calendar table
+## Regla crítica 1: siempre usar el Power BI MCP para escribir
 
-2. **Tables Analysis** (from `tables/*.tmdl`)
-   - For each table, extract:
-     - Columns (name, dataType, isHidden)
-     - Measures (name, expression, displayFolder)
-     - Calculated columns vs regular columns
-     - Partitions (data sources)
-   - Check naming conventions (PascalCase, spaces, prefixes)
+Cuando crees o modifiques medidas DAX, tablas, columnas, relaciones o cualquier elemento del modelo semántico, DEBES usar el servidor MCP de Power BI Desktop. NO edites archivos TMDL (`.tmdl`, `.tmd`) directamente. Esos archivos los gestiona el motor tabular. Editarlos a mano basado en tu conocimiento de entrenamiento es propenso a errores y probablemente rompa el modelo.
 
-3. **Relationships** (from `relationships.tmdl`)
-   - Count relationships
-   - Identify cardinality (one-to-many, many-to-many)
-   - Check for bi-directional filters (potential issues)
-   - Detect orphan tables (no relationships)
+## Regla crítica 2: verificar la conexión MCP al iniciar sesión
 
-4. **Measures Quality** (from table files)
-   - Count total measures
-   - Check for VAR/RETURN patterns (good)
-   - Check for nested CALCULATE (warning)
-   - Identify time intelligence usage
-   - Check for `formatStringDefinition` (dynamic formatting)
+Antes de hacer cualquier cambio al proyecto, verifica que el MCP de Power BI Desktop esté conectado a una instancia abierta. Si no lo está, avisa al usuario para que abra el archivo `.pbip` en Power BI Desktop primero.
 
-5. **Security** (from `roles/*.tmdl`)
-   - Count RLS roles
-   - Check filter expressions
+## Regla crítica 3: los archivos TMDL son SOLO LECTURA
 
-**TMDL Syntax Reference:**
-```tmdl
-table 'TableName'
-    column ColumnName
-        dataType: string
-        isHidden
-        lineageTag: ...
+Puedes LEER archivos TMDL para entender la estructura del modelo, las relaciones, las medidas. Es a veces más rápido que consultar por MCP. Pero NUNCA ESCRIBAS en esos archivos.
 
-    measure 'Measure Name' = SUM('Table'[Column])
-        displayFolder: "Folder"
-        formatStringDefinition = ...
-```
+## Regla crítica 4: consultar Microsoft Learn MCP para documentación técnica
 
-### For .pbip folders (Power BI Project format):
+Cuando el usuario pregunte sobre funciones DAX, propiedades de Power BI, conectores o capacidades de Fabric, consulta el Microsoft Learn MCP antes de responder. No respondas de memoria sobre temas donde la documentación puede haber cambiado desde tu fecha de corte de entrenamiento. Si el Learn MCP no está disponible, avisa al usuario antes de responder.
 
-```
-Analyzing Power BI Project: [name]
-==================================
+## Regla crítica 5: flujo completo
 
-I can read the project structure directly. Checking...
+Escribir via MCP. Leer via archivos. Versionar via Git. Después de cambios significativos, sugiere al usuario hacer un commit.
 
-📂 [name].pbip/
-   ├── 📄 definition.pbir (report definition)
-   ├── 📂 definition/ (report pages)
-   └── 📂 [name].SemanticModel/
-       ├── 📄 definition.bism
-       ├── 📄 model.bim (data model)
-       └── 📂 definition/
-           ├── tables/
-           ├── relationships/
-           └── expressions/
-```
+## Regla crítica 6: mantener ROADMAP.md al día
 
-**Analyze and report on:**
+Al final de cada sesión donde hayas hecho cambios al modelo, actualizá `./ROADMAP.md`:
+- Marcá con `[x]` las tareas que se completaron
+- Agregá una entrada a la sección "Bitácora" con la fecha, qué se hizo, y qué queda pendiente
+- Si detectaste tareas nuevas que no estaban listadas, agregalas a la fase correcta
 
-1. **Data Model Structure**
-   - Number of tables (fact vs dimension)
-   - Relationships (count, types, issues)
-   - Calculated columns vs measures ratio
+## Regla crítica 7: capturar aprendizajes en LEARNINGS.md
 
-2. **DAX Quality**
-   - Number of measures
-   - Naming conventions (prefixes, consistency)
-   - Complex measures identified
-   - Potential optimization opportunities
+Al inicio de cada sesión, LEÉ `./LEARNINGS.md` para no repetir errores ya documentados.
 
-3. **Power Query**
-   - Number of queries
-   - Data sources used
-   - Query folding potential
-   - Transformation patterns
+Al detectar cualquiera de estos casos, agregá una entrada nueva a `LEARNINGS.md`:
+- El usuario dice "esto no funcionó" / "acordate de X" / "la próxima vez usá..."
+- Un error técnico que tomó más de un intento arreglar
+- Un patrón específico del dominio o de las fuentes de datos que no es obvio desde el conocimiento general de Power BI
 
-4. **Report Structure**
-   - Number of pages
-   - Visuals per page (average)
-   - Theme applied (if any)
+Formato de cada entrada: **Fecha** — **Lección** — **Aplicación a futuro**.
 
-### For .pbix files:
+---
 
-```
-Note: .pbix files are compressed and require Power BI Desktop to fully analyze.
+## Estructura del proyecto
 
-I can see this file exists:
-📊 [filename.pbix] - [size] - Modified: [date]
+La carpeta raíz del proyecto es un repositorio Git independiente.
 
-For detailed analysis, I recommend:
-1. Convert to .pbip format (File > Save as > Power BI Project)
-2. Open in Power BI Desktop and connect via MCP
+- Carpeta `/pbip-files/`: contiene todos los archivos del proyecto Power BI
+- Archivo `/pbip-files/{projectName}.pbip`: archivo del proyecto Power BI
+- Carpeta `/pbip-files/{projectName}.SemanticModel/`: definición del modelo (archivos TMDL, **SOLO LECTURA para agentes**)
+- Carpeta `/pbip-files/{projectName}.Report/`: definición del reporte (**SOLO LECTURA para agentes**)
 
-Would you like guidance on converting to .pbip format?
+Cualquier otra carpeta en la raíz (`scripts/`, `docs/`, `notebooks/`, etc.) es espacio libre del proyecto y no debe mezclarse con los archivos de Power BI.
 ```
 
-### For Dataflows (dataflows/ folder with .pq files):
-
-```
-Analyzing Dataflows
-===================
+### 5.2 — Write `./CLAUDE.md` (pointer for Claude Code)
 
-I found Power Query dataflow definitions. Checking structure...
+```md
+<!-- Claude Code reads this file. The canonical instructions are in AGENTS.md. -->
 
-📂 dataflows/
-   ├── 📂 Customers/ ([X] queries)
-   ├── 📂 Products/ ([X] queries)
-   ├── 📂 Invoices/ ([X] queries)
-   └── ...
+Véase [AGENTS.md](./AGENTS.md) para las instrucciones de este proyecto.
 ```
 
-**Analyze .pq files and report on:**
+### 5.3 — Write `./.github/copilot-instructions.md` (pointer for GitHub Copilot)
 
-1. **Dataflow Organization**
-   - Count dataflow folders (each folder = one logical dataflow)
-   - Count queries per dataflow
-   - Identify naming patterns
+Create `./.github/` directory if it doesn't exist.
 
-2. **Query Analysis** (from `*.pq` files)
-   - For each query, extract:
-     - Query name (from filename)
-     - Data source type (BigQuery, SQL, API, etc.)
-     - Key transformations used
-   - Check for common patterns:
-     - `shared` keyword (reusable queries)
-     - Parameters usage
-     - Error handling (`try...otherwise`)
+```md
+<!-- GitHub Copilot reads this file. The canonical instructions are in AGENTS.md at the project root. -->
 
-3. **Data Sources**
-   - Identify unique data sources across all queries
-   - Check for connection strings / credentials references
-   - Map sources to dataflows
-
-4. **Quality Indicators**
-   - **Good**: Typed columns, proper naming, comments
-   - **Warning**: Hardcoded values, no error handling
-   - **Issue**: Circular references, disabled queries
-
-**Power Query (.pq) Syntax Reference:**
-```powerquery
-let
-    Source = GoogleBigQuery.Database("project"),
-    Data = Source{[Schema="schema", Item="table"]}[Data],
-    Filtered = Table.SelectRows(Data, each [Status] = "Active"),
-    Typed = Table.TransformColumnTypes(Filtered, {{"Date", type date}})
-in
-    Typed
+Véase [AGENTS.md](../AGENTS.md) para las instrucciones de este proyecto.
 ```
 
-### For .xlsx files:
-
-```
-Analyzing Excel file: [name.xlsx]
-=================================
+### 5.4 — Write `./.kilo/rules/project.md` (pointer for Kilo Code)
 
-Sheets found: [count]
-[List sheet names]
+Create `./.kilo/rules/` directory if it doesn't exist.
 
-I can analyze:
-1. Data structure and quality
-2. Formulas and calculations
-3. Named ranges and tables
-4. Potential for Power BI import
+```md
+<!-- Kilo Code reads files under .kilo/rules/. The canonical instructions are in AGENTS.md at the project root. -->
 
-Which aspect interests you most?
+Véase [AGENTS.md](../../AGENTS.md) para las instrucciones de este proyecto.
 ```
 
----
-
-## PHASE 4: Findings Report
+### 5.5 — Write `./GEMINI.md` (pointer for Gemini CLI)
 
-After analysis, present a structured report:
+```md
+<!-- Gemini CLI reads this file when configured. The canonical instructions are in AGENTS.md. -->
 
+Véase [AGENTS.md](./AGENTS.md) para las instrucciones de este proyecto.
 ```
-📋 PROJECT ANALYSIS REPORT
-==========================
-
-## Overview
-- Project type: [Power BI Desktop / PBIP / Fabric TMDL / Excel / Mixed]
-- Files analyzed: [count]
-- Overall health: [🟢 Good / 🟡 Needs attention / 🔴 Significant issues]
-
-## Data Model Assessment
-
-| Aspect | Status | Finding |
-|--------|--------|---------|
-| Star Schema | 🟢/🟡/🔴 | [Finding] |
-| Relationships | 🟢/🟡/🔴 | [Finding] |
-| Naming Conventions | 🟢/🟡/🔴 | [Finding] |
-| Calculated Columns | 🟢/🟡/🔴 | [Finding] |
-
-## DAX Assessment
 
-| Aspect | Status | Finding |
-|--------|--------|---------|
-| Measure Organization | 🟢/🟡/🔴 | [Finding] |
-| Naming Patterns | 🟢/🟡/🔴 | [Finding] |
-| Performance Patterns | 🟢/🟡/🔴 | [Finding] |
-| Time Intelligence | 🟢/🟡/🔴 | [Finding] |
+### 5.6 — Write `./ROADMAP.md`
 
-## Power Query Assessment
+Use the template below, filling in `{projectName}` and `{today}` (YYYY-MM-DD). Pick the fases body matching the domain captured in PHASE 4:
 
-| Aspect | Status | Finding |
-|--------|--------|---------|
-| Query Organization | 🟢/🟡/🔴 | [Finding] |
-| Data Types | 🟢/🟡/🔴 | [Finding] |
-| Query Folding | 🟢/🟡/🔴 | [Finding] |
-| Error Handling | 🟢/🟡/🔴 | [Finding] |
+```md
+# ROADMAP — {projectName}
 
-## Dataflows Assessment (if applicable)
+> Actualizado automáticamente por el agente al final de cada sesión.
+> Si algo está desactualizado, pedime "actualizá el ROADMAP".
 
-| Aspect | Status | Finding |
-|--------|--------|---------|
-| Organization | 🟢/🟡/🔴 | [X] dataflows, [Y] total queries |
-| Naming Conventions | 🟢/🟡/🔴 | [Finding] |
-| Data Sources | 🟢/🟡/🔴 | [List sources: BigQuery, API, etc.] |
-| Reusability | 🟢/🟡/🔴 | [Finding on shared queries] |
+## Fases planificadas
 
-## Security Assessment (if RLS present)
+{fases-for-domain}
 
-| Aspect | Status | Finding |
-|--------|--------|---------|
-| RLS Roles | 🟢/🟡/🔴 | [X] roles defined |
-| Filter Logic | 🟢/🟡/🔴 | [Finding] |
-| Coverage | 🟢/🟡/🔴 | [All sensitive tables covered?] |
-
-## Top Recommendations
+---
 
-1. **[Priority: High/Medium/Low]** [Recommendation]
-   - Impact: [Description]
-   - Effort: [Easy/Medium/Complex]
+## Bitácora
 
-2. **[Priority]** [Recommendation]
-   ...
+_(El agente agrega una entrada acá al final de cada sesión.)_
 
-(Showing top 5 recommendations)
+- **{today}** — Kickoff completado. Proyecto `{projectName}` creado con estructura `pbip-files/`. Próximo paso: empezar Fase 1.
 ```
 
----
+**Fases body by domain** — copy the block that matches:
 
-## PHASE 5: Action Plan
+**Finance**
 
-```
-Based on my analysis, here's a suggested action plan:
+```md
+### Fase 1: Modelo base
+- [ ] Tabla de hechos `GeneralLedger`
+- [ ] Dimensiones: `Date`, `Account`, `CostCenter`, `Scenario`
+- [ ] Marcar `Date` como tabla de fecha
+- [ ] Relaciones entre fact y dims
 
-## Quick Wins (do now)
-- [ ] [Action 1]
-- [ ] [Action 2]
+### Fase 2: Medidas core
+- [ ] Medidas base (sum, count, avg)
+- [ ] Time intelligence (YTD, YoY, MoM)
+- [ ] Variance vs. presupuesto
 
-## Short-term Improvements
-- [ ] [Action 3]
-- [ ] [Action 4]
-
-## Longer-term Enhancements
-- [ ] [Action 5]
-- [ ] [Action 6]
-
----
+### Fase 3: Seguridad y gobierno
+- [ ] RLS por CostCenter (si aplica)
+- [ ] Naming conventions documentadas en AGENTS.md
+- [ ] Display folders
 
-How would you like to proceed?
-
-1. 🔧 Start fixing issues - I'll guide you through improvements
-2. 📊 Redesign data model - Use the data-model-design wizard
-3. 🎨 Apply a theme - Use the theme-tweaker wizard
-4. 📝 Export this report - Save analysis as markdown
-5. 💬 Ask questions - Discuss specific findings
+### Fase 4: Reporte
+- [ ] Página resumen (KPIs principales)
+- [ ] Detalle por dimensión
+- [ ] Theme aplicado
 ```
 
----
+**Sales / Revenue**
 
-## PHASE 6: Migration Path
+```md
+### Fase 1: Modelo base
+- [ ] Tabla de hechos `Sales`
+- [ ] Dimensiones: `Date`, `Customer`, `Product`, `Salesperson`
+- [ ] Marcar `Date` como tabla de fecha
+- [ ] Relaciones entre fact y dims
 
-If user selected "Migration":
+### Fase 2: Medidas core
+- [ ] Revenue, units, AOV
+- [ ] Time intelligence (YTD, YoY, MoM)
+- [ ] Rankings (top N customers, top N products)
 
-```
-What are you migrating from?
+### Fase 3: Seguridad y gobierno
+- [ ] RLS por Salesperson / territorio (si aplica)
+- [ ] Naming conventions documentadas en AGENTS.md
+- [ ] Display folders
 
-1. Excel-only reporting → Power BI
-2. Other BI tool → Power BI (Tableau, Looker, etc.)
-3. Legacy Power BI → Modern Power BI (DirectQuery to Import, etc.)
-4. On-premise → Power BI Service (cloud)
+### Fase 4: Reporte
+- [ ] Página resumen (pipeline + cierre del mes)
+- [ ] Detalle por cliente / producto
+- [ ] Theme aplicado
 ```
 
-Then provide migration-specific guidance:
+**HR / People**
 
-### Excel → Power BI:
-```
-Excel to Power BI migration checklist:
+```md
+### Fase 1: Modelo base
+- [ ] Tabla de hechos `HeadcountSnapshot`
+- [ ] Dimensiones: `Date`, `Employee`, `Department`, `Position`
+- [ ] Marcar `Date` como tabla de fecha
+- [ ] Relaciones entre fact y dims
 
-📋 Assessment:
-- [ ] Identify all Excel reports in scope
-- [ ] Document data sources for each
-- [ ] List key calculations and KPIs
-- [ ] Identify report consumers and their needs
+### Fase 2: Medidas core
+- [ ] Headcount, altas, bajas, attrition rate
+- [ ] Time intelligence (MoM, YoY)
+- [ ] Compensación agregada por dim
 
-🔄 Data Layer:
-- [ ] Evaluate data refresh requirements
-- [ ] Design star schema for common data
-- [ ] Plan Power Query transformations
-- [ ] Consider dataflows for shared data
+### Fase 3: Seguridad y gobierno
+- [ ] RLS por Department (datos sensibles)
+- [ ] Naming conventions documentadas en AGENTS.md
+- [ ] Display folders
 
-📊 Report Layer:
-- [ ] Map Excel visuals to Power BI equivalents
-- [ ] Design navigation structure
-- [ ] Plan for interactivity (slicers, drill-through)
-- [ ] Consider mobile layouts
-
-Would you like to start with any of these areas?
+### Fase 4: Reporte
+- [ ] Página resumen (headcount + attrition)
+- [ ] Detalle por departamento
+- [ ] Theme aplicado
 ```
 
----
-
-## TOOLKIT OVERVIEW
-
-When user wants to explore the toolkit:
-
-```
-📚 BI-PROJECT-OS TOOLKIT OVERVIEW
-=================================
-
-## 🎯 Interactive Wizards (Skills)
-
-| Skill | Purpose | When to use |
-|-------|---------|-------------|
-| data-model-design | Design semantic models from scratch | New projects, major redesigns |
-| theme-tweaker | Customize Power BI themes | Branding, visual consistency |
-| project-kickoff | Analyze existing projects | You are here! |
-
-## 📖 Reference Skills
+**Operations / Supply Chain**
 
-| Skill | Purpose |
-|-------|---------|
-| dax | DAX patterns and best practices |
-| power-query | M language and transformations |
-| data-modeling | Star schema and relationships |
-| excel-formulas | Modern Excel formulas |
+```md
+### Fase 1: Modelo base
+- [ ] Tabla de hechos `Orders`
+- [ ] Dimensiones: `Date`, `Customer`, `Product`, `Facility`
+- [ ] Marcar `Date` como tabla de fecha
+- [ ] Relaciones entre fact y dims
 
-## 📝 Code Snippets
+### Fase 2: Medidas core
+- [ ] Volumen, SLA cumplido, órdenes abiertas
+- [ ] Time intelligence (YTD, rolling 4w)
+- [ ] Aging de órdenes
 
-| Category | Examples |
-|----------|----------|
-| DAX | Time intelligence, Rankings, CALCULATE patterns, KPIs |
-| Power Query | Data cleaning, Transformations |
-| Excel | XLOOKUP, Dynamic arrays |
+### Fase 3: Seguridad y gobierno
+- [ ] RLS por Facility (si aplica)
+- [ ] Naming conventions documentadas en AGENTS.md
+- [ ] Display folders
 
-## 🎨 Themes
-
-| Theme | Style |
-|-------|-------|
-| bi-superpowers-default | Shadcn/UI + IBCS standards |
-| corporate-blue | Professional blue palette |
-| dark-mode | Dark background |
+### Fase 4: Reporte
+- [ ] Página resumen (volumen + SLA)
+- [ ] Detalle por facility / producto
+- [ ] Theme aplicado
+```
 
-## 📁 Templates
+**Marketing**
 
-| Category | Contents |
-|----------|----------|
-| power-bi | Report templates |
-| excel | Workbook templates |
-| data-models | Model patterns |
+```md
+### Fase 1: Modelo base
+- [ ] Tabla de hechos `Campaigns`
+- [ ] Dimensiones: `Date`, `Channel`, `Audience`, `Product`
+- [ ] Marcar `Date` como tabla de fecha
+- [ ] Relaciones entre fact y dims
 
----
+### Fase 2: Medidas core
+- [ ] Spend, impressions, clicks, conversions, CPA
+- [ ] Time intelligence (YTD, rolling 4w)
+- [ ] Funnel rates (CTR, conversion rate)
 
-What would you like to explore?
+### Fase 3: Seguridad y gobierno
+- [ ] RLS por Audience / región (si aplica)
+- [ ] Naming conventions documentadas en AGENTS.md
+- [ ] Display folders
 
-1. See DAX snippets
-2. See Power Query snippets
-3. Preview available themes
-4. Learn about a specific skill
-5. Return to main menu
+### Fase 4: Reporte
+- [ ] Página resumen (spend + ROAS)
+- [ ] Detalle por canal / campaña
+- [ ] Theme aplicado
 ```
 
----
+**Otro (fallback)**
 
-## ANALYSIS PATTERNS
-
-### TMDL-Specific Patterns
-When analyzing `.tmdl` files:
-
-**File Structure:**
-- `table 'Name'` → Table definition starts
-- `column Name` → Column definition (indented with tab)
-- `measure 'Name' =` → Measure definition
-- `partition Name =` → Data source partition
-- `dataType:` → Column data type
-- `isHidden` → Hidden column/table flag
-- `lineageTag:` → Unique identifier (UUID)
-
-**Quality Checks:**
-- Measures should have `displayFolder` for organization
-- Dynamic formatting uses `formatStringDefinition` (not `formatString`)
-- Tables should have consistent column naming
-- Check for `isHidden` on technical columns
-
-**Common Issues:**
-- Missing `formatStringDefinition` for currency measures
-- Inconsistent indentation (must use tabs, not spaces)
-- Orphan tables without relationships
-- Measures without display folders
-
-### Star Schema Detection
-Check for:
-- Tables with "Dim" or "Dimension" prefix → Dimensions
-- Tables with "Fact" or transactional data → Facts
-- Tables with date ranges → Date dimension
-- Orphan tables (no relationships)
-- Bridge tables (many-to-many)
-
-### Naming Convention Analysis
-Check for:
-- Measure prefixes (e.g., "M_", "Measure_", or none)
-- Column naming (PascalCase, snake_case, spaces)
-- Table naming consistency
-- Hidden vs visible columns
-
-### DAX Quality Indicators
-- **Good**: CALCULATE with proper filters, variables (VAR/RETURN), time intelligence functions
-- **Warning**: Nested CALCULATE, FILTER on large tables, too many calculated columns
-- **Issue**: Circular dependencies, bi-directional relationships overuse
-
-### Power Query Quality Indicators
-- **Good**: Proper data types, query folding, organized steps
-- **Warning**: Many manual steps, hardcoded values, no error handling
-- **Issue**: Loading disabled queries, circular references
+```md
+### Fase 1: Modelo base
+- [ ] Definir tabla de hechos principal
+- [ ] Definir dimensiones clave (al menos `Date` + 2 más)
+- [ ] Marcar `Date` como tabla de fecha
+- [ ] Relaciones entre fact y dims
 
----
-
-## EXAMPLE CONVERSATIONS
+### Fase 2: Medidas core
+- [ ] Medidas base (sum, count, avg)
+- [ ] Time intelligence (YTD, YoY)
 
-### Example 1: Fabric TMDL Project
+### Fase 3: Seguridad y gobierno
+- [ ] RLS (si aplica al caso)
+- [ ] Naming conventions documentadas en AGENTS.md
+- [ ] Display folders
 
+### Fase 4: Reporte
+- [ ] Página resumen
+- [ ] Detalle por dimensión principal
+- [ ] Theme aplicado
 ```
-Agent: Welcome to BI Agent Superpowers! In which language...
-User: 2 (Español)
-
-Agent: Soy tu asistente de proyectos BI. ¿Es este un proyecto nuevo o tienes archivos existentes?
-User: 2 (Existente)
 
-Agent: Déjame escanear tu carpeta...
+### 5.7 — Write `./LEARNINGS.md`
 
-       RESULTADOS DEL ESCANEO
-       ======================
-       📁 Carpeta: /proyecto-fabric
+Use this seed template verbatim (solo cambiá `{projectName}`):
 
-       Encontré 284 archivos BI:
+```md
+# LEARNINGS — {projectName}
 
-       Fabric/TMDL:
-         🔷 data-model/definition/ - 113 archivos TMDL
-         📝 dataflows/ - 171 archivos Power Query
+> Lecciones aprendidas específicas de este proyecto.
+> El agente lee este archivo al inicio de cada sesión y agrega entradas cuando detecta algo que vale la pena recordar.
+>
+> Si querés agregar algo a mano, respetá el formato de abajo.
 
-       ¿Quieres que analice estos archivos en detalle?
+## Formato de cada entrada
 
-User: 1 (Sí, analizar todo)
+- **Fecha** (YYYY-MM-DD)
+- **Lección**: qué aprendimos
+- **Aplicación a futuro**: cómo usarlo la próxima vez
 
-Agent: Analizando modelo semántico Fabric...
+## Entradas
 
-       📂 data-model/definition/
-          ├── tables/ - 47 tablas
-          ├── roles/ - 8 roles RLS
-          └── relationships.tmdl - 35 relaciones
-
-       📋 REPORTE DE ANÁLISIS
-       ======================
-
-       Salud general: 🟢 Buena
+_(Vacío por ahora — las primeras entradas aparecerán cuando tengamos contexto que valga la pena guardar. El agente las agrega automáticamente.)_
+```
 
-       ## Modelo de Datos
-       | Aspecto | Estado | Hallazgo |
-       |---------|--------|----------|
-       | Star Schema | 🟢 | Estructura clara fact/dim |
-       | Relaciones | 🟢 | 35 relaciones, sin bi-direccionales |
-       | Naming | 🟡 | Algunas tablas con espacios |
-       | Medidas | 🟢 | 49 medidas organizadas en folders |
+### 5.8 — Confirm to the user
 
-       ## Dataflows
-       | Aspecto | Estado | Hallazgo |
-       |---------|--------|----------|
-       | Organización | 🟢 | 12 dataflows por dominio |
-       | Fuentes | 🟢 | BigQuery, Stripe, WooCommerce |
+```
+✓ Escribí:
+  • AGENTS.md                            ← documento canónico con las 7 reglas
+  • CLAUDE.md                            ← pointer para Claude Code
+  • .github/copilot-instructions.md      ← pointer para GitHub Copilot
+  • .kilo/rules/project.md               ← pointer para Kilo Code
+  • GEMINI.md                            ← pointer para Gemini CLI
+  • ROADMAP.md                           ← plan del proyecto (lo actualizo al final de cada sesión)
+  • LEARNINGS.md                         ← lecciones del proyecto (lo actualizo cuando algo valga la pena recordar)
 
-       ## Seguridad
-       | Aspecto | Estado | Hallazgo |
-       |---------|--------|----------|
-       | RLS Roles | 🟢 | 8 roles (uno por cliente) |
+A partir de ahora, cualquier agente que abra esta carpeta lee las mismas 7 reglas, el plan del proyecto, y los aprendizajes acumulados.
+```
 
-       ## Top Recomendaciones
-       1. [Media] Estandarizar nombres de tablas (quitar espacios)
-       2. [Baja] Agregar formatStringDefinition a medidas monetarias
+---
 
-       ¿Cómo quieres proceder?
+## PHASE 6: Start modeling
 
-User: 1 (Empezar a corregir)
+Now the foundation is ready. Propose the first concrete modeling step based on the domain captured in PHASE 4:
 
-Agent: Empecemos por la prioridad media: estandarizar nombres...
 ```
+Listo para modelar. Con lo que me contaste ({domain} / {purpose}), yo arrancaría por:
 
-### Example 2: Power BI Desktop Project
+1. Crear la tabla de hechos central — {suggested-fact-table-for-domain}
+2. Crear las dimensiones clave — {suggested-dims-for-domain}
+3. Conectar las fuentes de datos desde {dataSources[0]}
 
+¿Arrancamos por (1), o querés otro orden?
 ```
-Agent: Welcome to BI Agent Superpowers! In which language...
-User: 2 (Español)
 
-Agent: Soy tu asistente de proyectos BI. ¿Es este un proyecto nuevo o tienes archivos existentes?
-User: 2 (Existente)
+When the user picks, **execute the work via the Power BI Modeling MCP** — do not generate `.tmdl` text to paste. The user watches the model change live in Power BI Desktop.
 
-Agent: Déjame escanear tu carpeta...
+Suggested first steps by domain:
 
-       RESULTADOS DEL ESCANEO
-       ======================
-       Encontré 3 archivos:
-       📊 ventas-q4.pbip/ - Proyecto Power BI
-       📗 datos-maestros.xlsx - 2.3 MB
-       📗 presupuesto-2024.xlsx - 1.1 MB
+| Domain | Suggested fact table | Suggested dimensions |
+|---|---|---|
+| Finance | `GeneralLedger` | `Date`, `Account`, `CostCenter`, `Scenario` |
+| Sales | `Sales` | `Date`, `Customer`, `Product`, `Salesperson` |
+| HR | `HeadcountSnapshot` | `Date`, `Employee`, `Department`, `Position` |
+| Operations | `Orders` | `Date`, `Customer`, `Product`, `Facility` |
+| Marketing | `Campaigns` | `Date`, `Channel`, `Audience`, `Product` |
 
-       ¿Quieres que analice estos archivos en detalle?
+These are starting points, not gospel — adapt to what the user said in PHASE 4.
 
-User: 1 (Sí, analizar todo)
-
-Agent: [Analiza el .pbip]
-
-       📋 REPORTE DE ANÁLISIS
-       ======================
+---
 
-       Salud general: 🟡 Necesita atención
+## PHASE 7: Handoff to `/report-design`
 
-       ## Modelo de Datos
-       | Aspecto | Estado | Hallazgo |
-       |---------|--------|----------|
-       | Star Schema | 🟡 | 2 tablas sin relacionar |
-       | Relaciones | 🟢 | Bien configuradas |
-       | Naming | 🔴 | Inconsistente (mezcla español/inglés) |
+Once the model has at least 1 fact, 1 dim, and 3 measures in place, propose moving to reports:
 
-       ## Top Recomendaciones
-       1. [Alta] Estandarizar nombres de columnas
-       2. [Media] Crear tabla de fechas dedicada
-       3. [Baja] Mover cálculos de columnas a medidas
+```
+✓ Modelo base listo. Tenés fact + dims + {N} medidas.
 
-       ¿Cómo quieres proceder?
+El siguiente paso lógico es armar los 3 reportes con `/report-design`. Ese skill va a:
+  1. Usar el dominio que ya definimos ({domain})
+  2. Inspeccionar tu modelo vía el pbi CLI para las medidas/dimensiones exactas
+  3. Generar 3 páginas con visuales vía CLI (card, line, bar, matrix)
+  4. Cerrar y reabrir PBI Desktop para que renderice
 
-User: 1 (Empezar a corregir)
+Requisito: necesitás pbi-cli-tool instalado (si no lo tenés el skill te guía).
 
-Agent: Empecemos por la prioridad alta: estandarizar nombres...
+¿Arrancamos con /report-design, o preferís agregar más medidas al modelo primero?
 ```
 
+If the user opts for reports, load `/report-design` and let it run. If they want to keep modeling, stay on this skill.
+
 ---
 
-## Complexity Adaptation
+## What this skill does NOT do
 
-Adjust depth based on `config.json → experienceLevel`:
-- **beginner**: Step-by-step with explanations, reference library examples
-- **intermediate**: Standard depth, explain non-obvious decisions
-- **advanced**: Concise, skip basics, focus on edge cases and optimization
+- **No scoring / benchmarking** of an existing model. For that, the user can ask "audit this model" separately.
+- **Report authoring** is delegated to `/report-design` which orchestrates the `pbi-cli-tool` CLI — don't write `.Report/` files from here.
 
 ---
 
 ## Related Skills
 
-- `/data-model-design` — Design the star schema after analysis
-- `/migration-assistant` — If migration is needed
-- `/theme-tweaker` — Customize visuals after setup
-- `/governance` — Apply naming standards from the start
+- `/pbi-connect` — Deeper troubleshooting if the MCP connection fails
 
 ---
 
-## OUTPUT FILES
-
-When user requests to export the report, save to:
-
-`/[project-root]/analysis-report-[date].md`
+## Related Resources
 
-Format:
-```markdown
-# Project Analysis Report
-Generated: [date]
-Tool: BI Agent Superpowers
-
-## Executive Summary
-[Summary paragraph]
-
-## Files Analyzed
-[List]
-
-## Findings
-[Detailed findings by category]
-
-## Recommendations
-[Prioritized list]
-
-## Next Steps
-[Action items]
-```
+- [Power BI MCP overview](https://learn.microsoft.com/en-us/power-bi/developer/mcp/mcp-servers-overview)
+- [Power BI Modeling MCP on GitHub](https://github.com/microsoft/powerbi-modeling-mcp)
+- [PBIP format docs](https://learn.microsoft.com/en-us/power-bi/developer/projects/projects-overview)