@luquimbo/bi-superpowers 2.0.1 → 3.0.1

package/README.md

@@ -3,239 +3,122 @@
 [![npm version](https://img.shields.io/npm/v/@luquimbo/bi-superpowers)](https://www.npmjs.com/package/@luquimbo/bi-superpowers)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Node.js Version](https://img.shields.io/badge/node-%3E%3D18.0.0-brightgreen)](https://nodejs.org)
-[![Power BI](https://img.shields.io/badge/Power%20BI-Ready-F2C811?logo=powerbi)](https://powerbi.microsoft.com)
-[![Claude Code](https://img.shields.io/badge/Claude%20Code-plugin-191919)](https://docs.anthropic.com/en/docs/claude-code)
+[![Power BI](https://img.shields.io/badge/Power%20BI-Desktop-F2C811?logo=powerbi)](https://powerbi.microsoft.com)
 
-AI-powered toolkit for Power BI, Microsoft Fabric & Excel development. Ships as a native Claude Code plugin with 24 skills, 11 interactive commands, and MCP integration with official Microsoft servers.
+Toolkit open-source para trabajar con **Power BI Desktop** desde tu agente AI favorito. Funciona con los 5 agentes soportados: Claude Code, GitHub Copilot, Codex, Gemini CLI y Kilo Code.
 
-**Cross-platform**: Windows + macOS (Power BI Desktop and Excel MCP require Windows).
+> **Estado:** versión inicial. 2 skills y 2 MCP servers para empezar a iterar.
 
-## Works with
-
-| Platform | How | Status |
-|----------|-----|--------|
-| **Claude Code** | Native plugin (`--plugin-dir .`) | Full support |
-| **[1code.dev](https://1code.dev)** | Uses Claude Code SDK — plugins work automatically | Full support |
-| **Claude Desktop** | MCPB extension (`super build-desktop`) | Skills as MCP prompts |
+---
 
 ## Quick Start
 
-### Option 1 — Install skills for your AI agents (fastest)
-
 ```bash
-npx @luquimbo/bi-superpowers install
+# 1. Instalá la CLI
+npm install -g @luquimbo/bi-superpowers
+
+# 2. Instalá skills y MCPs en todos tus agentes
+super install
 ```
 
-Interactive installer — detects Claude Code, GitHub Copilot, Codex, Gemini CLI, Kilo Code and installs the 24 skills into the right directories.
+Eso es todo. El instalador:
 
-### Option 2 — Full CLI with plugin generation
+1. Detecta qué agentes AI tenés instalados (Claude Code, Copilot, Codex, Gemini, Kilo).
+2. Te deja elegir para cuáles instalar (o hace todos con `--all --yes`).
+3. Copia los 2 skills en los directorios correctos de cada agente.
+4. Escribe los 2 MCPs en el config file correcto de cada agente (con el formato/schema que cada uno espera).
 
-```bash
-# Install globally
-npm install -g @luquimbo/bi-superpowers
+Reiniciá tu agente AI, y ya tenés bi-superpowers funcionando.
 
-# Initialize the plugin in your project
-cd your-bi-project
-super kickoff
+---
 
-# Run with the plugin
-claude --plugin-dir .
-```
+## Qué viene incluido
 
-100% open source (MIT). No license keys. All 24 skills and the full library ship with the npm package.
+### 2 Skills
 
-### Claude Desktop
+| Skill | Para qué |
+|---|---|
+| `project-kickoff` | Analizá el estado de tu proyecto BI y decidí próximos pasos |
+| `pbi-connect` | Conectá tu agente AI a Power BI Desktop vía MCP |
 
-```bash
-super build-desktop
-# Double-click bi-superpowers.mcpb to install in Claude Desktop
-# 24 skills available as MCP prompts
-```
+### 2 MCP Servers (oficiales de Microsoft)
 
-### Free models via OpenRouter
+| MCP | Para qué |
+|---|---|
+| `powerbi-modeling` | Conexión local a Power BI Desktop (XMLA). Leer/escribir modelo semántico, ejecutar DAX. |
+| `microsoft-learn` | HTTP MCP de Microsoft Learn — docs oficiales de Power BI/DAX/Fabric en contexto de la conversación. |
 
-Using [1code.dev](https://1code.dev) with free AI models? See [docs/openrouter-free-models.md](./docs/openrouter-free-models.md) for setup instructions.
+---
 
-## Plugin Structure
+## Agentes soportados
 
-`super kickoff` generates the following in your project:
+| Agente | Skills | MCP config path |
+|---|---|---|
+| **Claude Code** | ✅ `~/.claude/skills/` | `~/.claude.json` |
+| **GitHub Copilot** | ✅ `~/.github/skills/` | `~/.copilot/mcp-config.json` |
+| **Codex (OpenAI)** | ✅ `~/.agents/skills/` | `~/.codex/config.toml` |
+| **Gemini CLI** | ✅ `~/.gemini/skills/` | `~/.gemini/settings.json` |
+| **Kilo Code** | ✅ `~/.kilocode/skills/` | `~/.kilocode/mcp_settings.json` |
 
-```
-your-project/
-├── .claude-plugin/plugin.json    # Plugin manifest
-├── .mcp.json                     # Power BI + Fabric MCPs
-├── commands/                     # 11 slash commands (interactive wizards)
-├── skills/                       # 24 skills (SKILL.md files)
-├── config.json                   # User preferences (experience level, language, domain)
-└── library/ → symlink            # Snippets, templates, themes
-```
+`super install` escribe todo esto automáticamente.
 
-## Skills (24)
-
-### Command Skills — 11 Interactive Wizards
-
-| Skill | Purpose | Trigger |
-|-------|---------|---------|
-| `/project-kickoff` | Analyze and plan BI projects | "new project", "analyze" |
-| `/data-model-design` | Interactive star schema builder | "design model" |
-| `/dax-doctor` | DAX debugging and optimization | "debug DAX", "fix measure" |
-| `/theme-tweaker` | Customize Power BI themes | "customize theme", "colors" |
-| `/pbi-connect` | Connect to Power BI Desktop | "connect Power BI" |
-| `/rls-design` | Row-level security design | "RLS", "security" |
-| `/fabric-scripts` | Microsoft Fabric automation | "Fabric scripts" |
-| `/model-documenter` | Semantic model documentation | "document model" |
-| `/migration-assistant` | Migration and upgrade wizard | "migrate", "move to Fabric" |
-| `/report-layout` | Report page layout planner | "report layout" |
-| `/contributions` | Contribution guidelines | "contribute", "PR" |
-
-### Reference Skills — 13 Background Knowledge
-
-| Skill | Purpose | Trigger |
-|-------|---------|---------|
-| `/dax` | DAX writing best practices | "DAX", "measure" |
-| `/power-query` | Power Query / M language | "Power Query", "M code" |
-| `/data-modeling` | Star schema theory and patterns | "star schema" |
-| `/excel-formulas` | Modern Excel formulas | "XLOOKUP", "Excel formula" |
-| `/fast-standard` | FP&A financial modeling (FAST) | "FAST standard" |
-| `/query-performance` | Performance optimization | "optimize", "slow" |
-| `/data-quality` | Data profiling and validation | "data quality" |
-| `/testing-validation` | DAX and data testing workflows | "testing", "test cases" |
-| `/governance` | Naming conventions and standards | "naming convention" |
-| `/semantic-model` | Semantic model best practices | "TMDL", "DirectLake" |
-| `/report-design` | Report design and visualization | "chart type", "IBCS" |
-| `/deployment` | CI/CD and deployment patterns | "CI/CD", "pipeline" |
-| `/dax-udf` | DAX user-defined functions | "UDF", "DEFINE FUNCTION" |
-
-## MCP Integration
-
-4 official Microsoft MCP servers configured automatically via `.mcp.json`:
-
-| MCP Server | Transport | Platform | Description |
-|------------|-----------|----------|-------------|
-| `powerbi-remote` | HTTP | Windows, macOS, Linux | Power BI remote operations via Fabric API |
-| `fabric-mcp-server` | stdio (`npx`) | Windows, macOS, Linux | Microsoft Fabric automation |
-| `powerbi-modeling-mcp` | stdio (local) | Windows only | Local Power BI Desktop modeling |
-| `microsoft-learn` | HTTP | Windows, macOS, Linux | Search Microsoft docs, code samples |
-
-> **Note**: `powerbi-modeling-mcp` requires Power BI Desktop (Windows only). `microsoft-learn` is free, requires no authentication, and provides access to the full Microsoft Learn documentation library.
-
-## CLI Commands
-
-| Command | Description |
-|---------|-------------|
-| `super kickoff` | Generate the Claude Code plugin in your project |
-| `super recharge` | Regenerate plugin after editing source skills |
-| `super build-desktop` | Build MCPB extension for Claude Desktop |
-| `super mcp-setup` | Configure MCP servers |
-| `super install` | Install skills into your AI agents' directories |
-| `super powers` | Show available skills, snippets, and themes |
-| `super upgrade` | Update to the latest version |
-| `super xray <query>` | Search snippets and library content |
-| `super checkup [file]` | Validate skill files |
-| `super scan` | Show diff between source and generated configs |
-| `super sentinel` | Watch for changes and auto-regenerate |
-| `super help` | Show help menu |
-
-## Library
-
-The `library/` folder contains reusable assets accessible from your prompts:
+### Instalación selectiva
 
-```
-library/
-├── snippets/           # Code patterns
-│   ├── dax/            # Time intelligence, CALCULATE, rankings, UDFs
-│   ├── power-query/    # Transformations, error handling, API integration
-│   ├── excel-formulas/ # Dynamic arrays, lookups, text functions
-│   ├── report-design/  # Chart selection, accessibility, layouts
-│   └── governance/     # Naming conventions, review checklists
-├── templates/          # Project starters (Power BI, Excel, data models)
-├── themes/             # Power BI JSON themes (6 themes included)
-└── examples/           # Reference implementations
-    ├── finance-reporting/
-    ├── hr-analytics/
-    ├── retail-analytics/
-    ├── supply-chain/
-    └── marketing-analytics/
+```bash
+# Solo para Claude Code
+super install --agent claude-code --yes
+
+# Solo para Claude Code y GitHub Copilot
+super install --agent claude-code --agent github-copilot --yes
+
+# Todos sin prompts
+super install --all --yes
 ```
 
-Use in prompts: *"Using patterns from library/snippets/dax/time-intelligence.md, create a rolling 12-month average measure"*
+---
 
-## Configuration
+## Actualizar
 
-`config.json` provides context to the AI about your experience level and domain:
+```bash
+# Actualizar la CLI a la última versión
+npm install -g @luquimbo/bi-superpowers@latest
 
-```json
-{
-  "experienceLevel": "intermediate",
-  "language": "en",
-  "workArea": "finance",
-  "powerBIConnection": "desktop"
-}
+# Re-instalar skills y MCPs (sobrescribe los viejos)
+super install --all --yes
 ```
 
-## Project Types
+---
 
-| Feature | Desktop (.pbix) | PBIP | Fabric |
-|---------|-----------------|------|--------|
-| Single file | Yes | No | N/A (cloud) |
-| Git-friendly | No | Yes | Yes |
-| Code reviews | No | Yes | Yes |
-| Works offline | Yes | Yes | No |
-| Team collaboration | Limited | Excellent | Excellent |
-| AI file access | Via running PBI | Direct file access | Via API |
-| CI/CD ready | No | Yes | Yes |
+## Requisitos
 
-## Repository Structure
+- **Node.js ≥ 18**
+- **Power BI Desktop en Windows** (para el MCP `powerbi-modeling` — el MCP `microsoft-learn` funciona en cualquier plataforma)
+- Al menos uno de: Claude Code, GitHub Copilot, Codex, Gemini CLI, Kilo Code
 
-```
-bi-superpowers/
-├── bin/                        # CLI tool and commands
-├── src/content/                # Source content (single source of truth)
-│   ├── base.md                 # Core AI identity
-│   ├── routing.md              # Skill activation logic
-│   └── skills/                 # 24 skill definitions
-├── library/                    # Reusable assets (snippets, templates, themes)
-├── desktop-extension/          # Claude Desktop MCPB template
-├── docs/                       # Additional documentation
-├── commands/                   # Generated slash commands
-├── skills/                     # Generated SKILL.md files
-├── .claude-plugin/             # Plugin manifest
-├── .mcp.json                   # MCP server configuration
-├── config.json                 # AI preferences
-└── AGENTS.md                   # AI context and documentation
-```
+---
 
-## Development
+## Comandos
 
 ```bash
-npm install
-npm run build:plugin    # Generate 11 commands + 24 skills
-npm test                # 77 tests
-npm run check           # ESLint + Prettier
+super install          # Instalá skills + MCPs en tus agentes AI
+super kickoff          # Generá el plugin de Claude Code en un proyecto (avanzado)
+super recharge         # Regenerá el plugin tras editar skills fuente
+super powers           # Listá skills disponibles
+super upgrade          # Actualizá la CLI a la última versión
+super help             # Ayuda completa
 ```
 
-## Resources
-
-- [Claude Code plugins](https://docs.anthropic.com/en/docs/claude-code)
-- [Power BI MCP overview](https://learn.microsoft.com/en-us/power-bi/developer/mcp/mcp-servers-overview)
-- [Microsoft Fabric MCP](https://github.com/microsoft/mcp/tree/main/servers/Fabric.Mcp.Server)
-- [DAX Guide](https://dax.guide)
-- [OpenRouter free models](./docs/openrouter-free-models.md)
+---
 
-## Author
+## Contribuir
 
-**Lucas Sánchez** — BI Consultant with 10+ years delivering Power BI, Excel, and data analytics projects.
+Este proyecto está empezando. Si querés agregar skills, mejorar los existentes, o reportar bugs:
 
-Creator of **[Acadevor.com](https://acadevor.com)**, the leading Spanish-speaking community for Business Intelligence professionals.
+- Abrí un issue en https://github.com/luquimbo/bi-superpowers/issues
+- Pull requests bienvenidas
 
-- [LinkedIn](https://www.linkedin.com/in/luquimbo/)
-- [GitHub](https://github.com/luquimbo)
-- [Acadevor](https://acadevor.com)
+---
 
 ## License
 
-[MIT](./LICENSE)
-
----
-
-**BI Agent Superpowers** v1.0.0 — AI-powered toolkit for Business Intelligence.
+MIT © Lucas Sanchez ([@luquimbo](https://github.com/luquimbo))

package/bin/cli.js

@@ -164,57 +164,56 @@ BI Agent Superpowers v${VERSION}
 ================================
 
 Claude Code plugin for Power BI, Fabric & Excel development.
-Works with: Claude Code, 1code.dev, Claude Desktop.
+
+Quick Start:
+  super kickoff                  # Genera el plugin en tu proyecto
+  claude --plugin-dir .          # Abrí Claude Code con el plugin
 
 Usage:
   super <command> [options]
 
-Commands:
-  install           Install skills for your AI agents (Claude, Copilot, Codex...)
-  kickoff [path]    Initialize the Claude Code plugin in your project
-  recharge [path]   Regenerate the plugin from source skills
-  build-desktop     Build .mcpb extension for Claude Desktop
-  powers            List available skills and resources
-  xray <query>      Search snippets and library content
-  checkup [file]    Validate skill files
-  scan              Show changes between source and generated configs
-  sentinel          Watch for changes and auto-regenerate
-  mcp-setup         Configure official Microsoft MCP servers
-  upgrade           Update to latest version
-  about             Show installation info
-  help              Show this help
-
-Repo Multi-Proyecto (v3):
-  setup           Create your bi-repo for version control
-  add <file>      Add a .pbix/.xlsx project to the repo
-  pull [project]  Pull changes from original file to repo
-  push [project]  Push changes from repo to original file
-  sync-source     Bidirectional sync (detects which is newer)
-  sync-profile    Sync snippets to a base profile
+Primary commands (Claude Code):
+  kickoff [path]    Genera el plugin completo (skills + commands + MCPs)
+  recharge [path]   Regenerá el plugin tras editar skills fuente
+  build-desktop     Buildea la extensión .mcpb para Claude Desktop
+  mcp-setup         Configurá los MCP servers de Microsoft
+  powers            Listá skills y recursos disponibles
+  upgrade           Actualizá a la última versión
+  about             Info de la instalación
+  help              Mostrá esta ayuda
+
+Developer tools:
+  xray <query>      Buscá snippets y contenido de la biblioteca
+  checkup [file]    Validá archivos de skills
+  scan              Diff entre source y generated
+  sentinel          Watch y auto-regenerá
 
-Options:
-  --dry-run       Preview changes without creating files (kickoff, recharge)
+Repo Multi-Proyecto:
+  setup             Creá tu bi-repo para version control
+  add <file>        Agregá un proyecto .pbix/.xlsx al repo
+  pull [project]    Pull changes desde el archivo original
+  push [project]    Push changes al archivo original
+  sync-source       Bidirectional sync (detecta cuál es más nuevo)
+  sync-profile      Sync snippets al profile base
 
-Install options:
-  --agent, -a     Specify agents (e.g. -a claude-code -a codex)
-  --all           Install for all supported agents
-  --yes, -y       Skip confirmation prompts
+Experimental (solo skills, sin commands ni MCPs):
+  install           Instalá skills en otros agentes AI (Copilot, Codex, Gemini, Kilo)
+
+Options:
+  --dry-run         Preview de cambios sin escribir archivos (kickoff, recharge)
 
 Examples:
-  super install                  # Interactive multi-agent installer
-  super install -a claude-code   # Install for Claude Code only
-  super install --all --yes      # Install for all agents, no prompts
-  super kickoff                  # Initialize plugin in current directory
-  super kickoff ./my-project     # Initialize in specific directory
-  super kickoff --dry-run        # Preview what would be created
-  super recharge                 # Regenerate plugin from source skills
-  super build-desktop            # Build .mcpb for Claude Desktop
-  super xray "YTD"               # Search for time intelligence patterns
-  super xray --category dax      # Search within a category
-  super checkup                  # Validate all skills
-  super mcp-setup                # Configure official Microsoft MCPs
-  super powers                   # Show available superpowers
-  claude --plugin-dir .          # Run Claude Code with the plugin
+  super kickoff                  # Inicializá el plugin en el directorio actual
+  super kickoff ./my-project     # Inicializá en un directorio específico
+  super kickoff --dry-run        # Preview de qué archivos se crearían
+  super recharge                 # Regenerá el plugin tras editar skills
+  super build-desktop            # Buildeá .mcpb para Claude Desktop
+  super mcp-setup                # Configurá los MCPs de Microsoft
+  super powers                   # Mostrá todos los skills disponibles
+  claude --plugin-dir .          # Corré Claude Code con el plugin
+
+  super install                  # [Experimental] Instalá skills en otros agentes
+  super install -a claude-code   # Instalá solo para un agente específico
 
 Repo Multi-Proyecto:
   super setup                    # Create your bi-repo (first time)

package/bin/commands/install.js

@@ -26,6 +26,7 @@ const path = require('path');
 const os = require('os');
 const readline = require('readline');
 const { AGENTS, UNIVERSAL_DIR } = require('../lib/agents');
+const { writeMcpConfigForAgent } = require('../lib/mcp-config');
 
 /**
  * Detecta qué agentes están instalados revisando sus directorios de config.
@@ -348,7 +349,7 @@ async function installCommand(args, config) {
     )
   );
 
-  // Instalar
+  // Instalar skills
   let agentResults;
   let copyFallbacks;
   try {
@@ -356,18 +357,18 @@ async function installCommand(args, config) {
     agentResults = result.agentResults;
     copyFallbacks = result.copyFallbacks;
   } catch (err) {
-    console.error(chalk.red(formatFsError(err, 'Falló la instalación')));
+    console.error(chalk.red(formatFsError(err, 'Falló la instalación de skills')));
     process.exit(1);
   }
 
-  // Universal
+  // Universal skills path
   const universalAgents = selectedAgents
     .filter((id) => AGENTS[id] && AGENTS[id].dir === UNIVERSAL_DIR)
     .map((id) => AGENTS[id].name);
   const universalSuffix = universalAgents.length > 0 ? ` — ${universalAgents.join(', ')}` : '';
   console.log(chalk.green(`  ✓ ${UNIVERSAL_DIR}/ (${skillDirs.length} skills)${universalSuffix}`));
 
-  // Por agente
+  // Skills por agente
   for (const result of agentResults) {
     const icon = result.method === 'symlinked' ? '→' : '✓';
     console.log(chalk.green(`  ${icon} ${result.dir}/ (${result.method}) — ${result.agent}`));
@@ -384,23 +385,73 @@ async function installCommand(args, config) {
     );
   }
 
+  // Configurar MCPs para cada agente seleccionado
+  console.log(
+    chalk.cyan('\n  Configurando MCP servers (Power BI Modeling + Microsoft Learn)...\n')
+  );
+  const mcpResults = [];
+  for (const agentId of selectedAgents) {
+    const agent = AGENTS[agentId];
+    try {
+      const configPath = writeMcpConfigForAgent(agentId, packageDir);
+      if (configPath) {
+        const relPath = configPath.replace(baseDir, '~');
+        console.log(chalk.green(`  ✓ ${relPath} — ${agent.name}`));
+        mcpResults.push({ agent: agent.name, configPath, success: true });
+      }
+    } catch (err) {
+      console.log(chalk.red(`  ✗ ${agent.name}: ${err.message}`));
+      mcpResults.push({ agent: agent.name, success: false, error: err.message });
+    }
+  }
+
   // Resumen
   const totalAgents = agentResults.length + (universalAgents.length > 0 ? 1 : 0);
+  const mcpSuccess = mcpResults.filter((r) => r.success).length;
+  const mcpFailures = mcpResults.filter((r) => !r.success);
+  const hasFailures = mcpFailures.length > 0;
+
+  const successMsg = `Instalados ${skillDirs.length} skills + 2 MCPs para ${totalAgents} agentes`;
+  const failureMsg = `Instalados ${skillDirs.length} skills. MCPs: ${mcpSuccess}/${mcpResults.length} agentes ✓, ${mcpFailures.length} con errores.`;
+  const headerLine = hasFailures ? chalk.yellow.bold(failureMsg) : chalk.green.bold(successMsg);
+
+  const failureDetail = hasFailures
+    ? '\n' +
+      chalk.red('Agentes con errores en MCP:') +
+      '\n' +
+      mcpFailures.map((r) => chalk.red(`  ✗ ${r.agent}: ${r.error}`)).join('\n') +
+      '\n'
+    : '';
+
   console.log(
     boxen(
-      chalk.green.bold(`Instalados ${skillDirs.length} skills para ${totalAgents} agentes`) +
+      headerLine +
+        failureDetail +
+        '\n\n' +
+        chalk.gray(`MCPs configurados en ${mcpSuccess}/${mcpResults.length} agentes.`) +
+        '\n' +
+        chalk.gray('Reiniciá tu agente AI para que tome los MCPs nuevos.') +
         '\n\n' +
-        chalk.gray('Los skills están listos. Abrí tu agente AI y empezá a usarlos.') +
+        chalk.gray('Los 2 skills disponibles:') +
         '\n' +
-        chalk.gray('Ejemplo: "Ayudame a escribir una medida DAX de ventas YTD"'),
+        chalk.gray('  /project-kickoff  — Analizá tu proyecto BI') +
+        '\n' +
+        chalk.gray('  /pbi-connect      — Conectá Claude a Power BI Desktop'),
       {
         padding: 1,
         margin: { top: 1 },
         borderStyle: 'round',
-        borderColor: 'green',
+        borderColor: hasFailures ? 'yellow' : 'green',
       }
     )
   );
+
+  if (hasFailures) {
+    // Non-zero exit so CI/scripts know something went wrong, but skills
+    // still got installed — we use exit code 2 to distinguish from total
+    // failure (exit 1).
+    process.exitCode = 2;
+  }
 }
 
 // Exports internos para testing

package/bin/commands/install.test.js

@@ -210,3 +210,80 @@ describe('install command - module exports', () => {
     assert.strictEqual(typeof installCommand.formatFsError, 'function');
   });
 });
+
+describe('install command - integration: --all --yes', () => {
+  let tempHome;
+  let tempPkg;
+  let origHome;
+
+  beforeEach(() => {
+    tempHome = fs.mkdtempSync(path.join(os.tmpdir(), 'bi-install-int-home-'));
+    tempPkg = fs.mkdtempSync(path.join(os.tmpdir(), 'bi-install-int-pkg-'));
+
+    // Minimal fake package layout: skills/<name>/SKILL.md + launcher file
+    const skillsDir = path.join(tempPkg, 'skills');
+    for (const skillName of ['project-kickoff', 'pbi-connect']) {
+      const dir = path.join(skillsDir, skillName);
+      fs.mkdirSync(dir, { recursive: true });
+      fs.writeFileSync(
+        path.join(dir, 'SKILL.md'),
+        `---\nname: ${skillName}\ndescription: fake skill for test\n---\n# ${skillName}\n`
+      );
+    }
+    fs.mkdirSync(path.join(tempPkg, 'bin', 'mcp'), { recursive: true });
+    fs.writeFileSync(
+      path.join(tempPkg, 'bin', 'mcp', 'powerbi-modeling-launcher.js'),
+      '// fake launcher\n'
+    );
+
+    origHome = os.homedir;
+    os.homedir = () => tempHome;
+  });
+
+  afterEach(() => {
+    os.homedir = origHome;
+    fs.rmSync(tempHome, { recursive: true, force: true });
+    fs.rmSync(tempPkg, { recursive: true, force: true });
+  });
+
+  test('installs skills and writes MCP config for all 5 agents', async () => {
+    const origExitCode = process.exitCode;
+    try {
+      await installCommand(['--all', '--yes'], {
+        packageDir: tempPkg,
+        version: '9.9.9-test',
+      });
+    } finally {
+      // Reset exit code so subsequent tests aren't affected
+      process.exitCode = origExitCode;
+    }
+
+    // Skills installed at universal path
+    assert.ok(
+      fs.existsSync(path.join(tempHome, '.agents', 'skills', 'project-kickoff', 'SKILL.md'))
+    );
+    assert.ok(fs.existsSync(path.join(tempHome, '.agents', 'skills', 'pbi-connect', 'SKILL.md')));
+
+    // MCP configs written for all 5 agents
+    const expectedMcpFiles = [
+      path.join(tempHome, '.claude.json'),
+      path.join(tempHome, '.copilot', 'mcp-config.json'),
+      path.join(tempHome, '.codex', 'config.toml'),
+      path.join(tempHome, '.gemini', 'settings.json'),
+      path.join(tempHome, '.kilocode', 'mcp_settings.json'),
+    ];
+    for (const filePath of expectedMcpFiles) {
+      assert.ok(fs.existsSync(filePath), `expected MCP config at ${filePath}`);
+    }
+
+    // Claude Code config has both servers under mcpServers
+    const claudeConfig = JSON.parse(fs.readFileSync(path.join(tempHome, '.claude.json'), 'utf8'));
+    assert.ok(claudeConfig.mcpServers['powerbi-modeling']);
+    assert.ok(claudeConfig.mcpServers['microsoft-learn']);
+
+    // Codex TOML has both sections
+    const codexToml = fs.readFileSync(path.join(tempHome, '.codex', 'config.toml'), 'utf8');
+    assert.ok(codexToml.includes('[mcp_servers.powerbi-modeling]'));
+    assert.ok(codexToml.includes('[mcp_servers.microsoft-learn]'));
+  });
+});

package/bin/lib/generators/claude-plugin.js

@@ -20,37 +20,12 @@ const {
 } = require('../microsoft-mcp');
 const { parseSkillMetadata, getSkillPurpose } = require('./shared');
 
-const COMMAND_SKILLS = new Set([
-  'project-kickoff',
-  'data-model-design',
-  'theme-tweaker',
-  'pbi-connect',
-  'rls-design',
-  'fabric-scripts',
-  'contributions',
-  // Phase 2 — new interactive wizards
-  'dax-doctor',
-  'model-documenter',
-  'migration-assistant',
-  'report-layout',
-]);
-
-const REFERENCE_SKILLS = new Set([
-  'dax',
-  'power-query',
-  'data-modeling',
-  'query-performance',
-  'testing-validation',
-  'data-quality',
-  'fast-standard',
-  'excel-formulas',
-  // Phase 3 — new background knowledge
-  'governance',
-  'semantic-model',
-  'report-design',
-  'deployment',
-  'dax-udf',
-]);
+// Currently the plugin ships 2 skills, both interactive wizards.
+// The old reference-skills split is kept as an empty set so that any
+// downstream code that checks `REFERENCE_SKILLS.has(x)` still works.
+const COMMAND_SKILLS = new Set(['project-kickoff', 'pbi-connect']);
+
+const REFERENCE_SKILLS = new Set();
 
 /**
  * Ensure a directory exists.