claude-autopm 1.26.0 → 1.28.0

package/README.md

@@ -44,6 +44,46 @@ PRD → Epic Decomposition → Parallel Development → Testing → Production
 
 ## ✨ Key Features
 
+### 🆕 **NEW in v1.28.0: Templates & Scaffolding!**
+
+**PRD Templates (Quick Start)**
+- 📋 **5 Built-in Templates** - api-feature, ui-feature, bug-fix, data-migration, documentation
+- 🚀 **70% Faster** - Create PRDs in 9 minutes instead of 30
+- 🎯 **Context7-Verified** - All templates use 2025 best practices
+- ✨ **Smart Variables** - Auto-generate IDs, timestamps, authors
+- 🎨 **Custom Templates** - Create your own team-specific templates
+
+**Template Commands**
+```bash
+# Create PRD from template
+autopm prd:new --template api-feature "Payment API"
+
+# Interactive template selection
+autopm prd:new "my-feature"
+
+# List available templates
+autopm template:list
+
+# Create custom template
+autopm template:new prd my-custom-template
+```
+
+**260+ Tests Passing** - Production-ready with comprehensive test coverage!
+
+---
+
+### 🎉 **v1.27.0: Phase 2 Complete!**
+
+**GitHub Sync (Bidirectional)**
+- 📤 **Upload to GitHub Issues** - Sync PRDs/Epics/Tasks with smart conflict detection
+- 📥 **Download from GitHub** - Pull Issues back to local files with reverse mapping
+- 🔄 **Bidirectional Mapping** - Maintain consistency with `sync-map.json`
+
+**Task Management**
+- ✅ **Complete Task Lifecycle** - List, show, update tasks within epics
+- 🔗 **Dependency Tracking** - Validate task dependencies automatically
+- 📊 **Progress Auto-update** - Epic progress updates on task completion
+
 ### 🤖 **39 Specialized AI Agents**
 
 Organized into dynamic teams that load based on your work context:

package/autopm/.claude/agents/frameworks/e2e-test-engineer.md

@@ -39,24 +39,7 @@ Access E2E testing frameworks and patterns:
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your E2E testing expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific E2E test framework and scenario guidance
-- Dependency and parallelization mapping
-- Test automation, visual regression, and accessibility testing considerations
+@include includes/task-creation-excellence.md
 
 ## E2E Testing Methodology
 

package/autopm/.claude/agents/frameworks/nats-messaging-expert.md

@@ -37,24 +37,7 @@ Use this agent for NATS messaging system including pub/sub, request/reply, and q
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your NATS messaging expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific NATS messaging pattern and JetStream guidance
-- Dependency and parallelization mapping
-- Pub/sub, clustering, and high availability considerations
+@include includes/task-creation-excellence.md
 
 ## Core Expertise
 

package/autopm/.claude/agents/frameworks/react-frontend-engineer.md

@@ -45,24 +45,7 @@ Before starting any implementation, you have access to live documentation throug
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your React expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific React implementation guidance
-- Dependency and parallelization mapping
-- Component architecture and state management considerations
+@include includes/task-creation-excellence.md
 
 **Core Expertise:**
 

package/autopm/.claude/agents/frameworks/react-ui-expert.md

@@ -32,24 +32,7 @@ Before implementing any UI solution, access live documentation through context7:
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your React UI expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific React UI framework implementation guidance
-- Dependency and parallelization mapping
-- Component design system and accessibility considerations
+@include includes/task-creation-excellence.md
 
 ## Test-Driven Development (TDD) Methodology
 

package/autopm/.claude/agents/frameworks/tailwindcss-expert.md

@@ -40,24 +40,7 @@ Access Tailwind CSS documentation through context7:
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your TailwindCSS expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific TailwindCSS utility and design system guidance
-- Dependency and parallelization mapping
-- Responsive design and theme customization considerations
+@include includes/task-creation-excellence.md
 
 ## When to Use This Agent
 

package/autopm/.claude/agents/frameworks/ux-design-expert.md

@@ -41,24 +41,7 @@ Access UX design and accessibility documentation:
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your UX design expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific UX/UI design and accessibility guidance
-- Dependency and parallelization mapping
-- User research, interaction design, and accessibility compliance considerations
+@include includes/task-creation-excellence.md
 
 ## Core Expertise
 

package/autopm/.claude/agents/languages/bash-scripting-expert.md

@@ -50,24 +50,7 @@ Before starting any implementation, you have access to live documentation throug
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development (with bats for Bash)
-4. **Technology-Specific Details**: Leverage your Bash/scripting expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific shell scripting implementation guidance
-- Dependency and parallelization mapping
-- Security and portability considerations included
+@include includes/task-creation-excellence.md
 
 ## Core Expertise
 

package/autopm/.claude/agents/languages/javascript-frontend-engineer.md

@@ -48,24 +48,7 @@ Before starting any implementation, you have access to live documentation throug
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your JavaScript/TypeScript expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific technical implementation guidance
-- Dependency and parallelization mapping
-- No vague descriptions or missing acceptance criteria
+@include includes/task-creation-excellence.md
 
 ## Core Expertise
 

package/autopm/.claude/agents/languages/nodejs-backend-engineer.md

@@ -52,24 +52,7 @@ Before starting any implementation, you have access to live documentation throug
 3. **Security Standards**: Access current security recommendations
 4. **Performance Patterns**: Get latest Node.js optimization techniques
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your Node.js/backend expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific technical implementation guidance for backend/APIs
-- Dependency and parallelization mapping
-- Security and performance considerations included
+@include includes/task-creation-excellence.md
 
 ## Core Expertise
 

package/autopm/.claude/agents/languages/python-backend-engineer.md

@@ -45,24 +45,7 @@ Before starting any implementation, you have access to live documentation throug
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your Python backend expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific Python/FastAPI implementation guidance
-- Dependency and parallelization mapping
-- API design, database integration, and async pattern considerations
+@include includes/task-creation-excellence.md
 
 **Core Expertise:**
 

package/autopm/.claude/agents/languages/python-backend-expert.md

@@ -32,24 +32,7 @@ Before implementing any Python backend solution, access live documentation throu
 - `mcp://context7/agile/acceptance-criteria` - Writing effective AC
 - `mcp://context7/project-management/estimation` - Effort estimation
 
-## Task Creation Excellence
-
-When creating implementation tasks (for epic decomposition or project planning):
-
-1. **Query Context7 First**: Access latest agile/PM best practices
-2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
-3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
-4. **Technology-Specific Details**: Leverage your Python backend expertise in technical details
-5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
-
-**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
-
-**Quality Standards**:
-- Tasks completable in 1-3 days maximum
-- Action-oriented titles (verb + noun)
-- Specific Python framework implementation guidance
-- Dependency and parallelization mapping
-- Architecture patterns, ORM integration, and async support considerations
+@include includes/task-creation-excellence.md
 
 ## Test-Driven Development (TDD) Methodology
 

package/autopm/.claude/commands/pm/epic-decompose.md

@@ -10,13 +10,13 @@ Break epic into concrete, actionable tasks.
 
 **Single Epic:**
 ```bash
-/pm:epic-decompose <feature_name>
+/pm:epic-decompose <feature_name> [--local]
 ```
 
 **Multi-Epic (decomposes ALL epics at once):**
 ```bash
-/pm:epic-decompose <feature_name>
-# Example: /pm:epic-decompose ecommerce-platform
+/pm:epic-decompose <feature_name> [--local]
+# Example: /pm:epic-decompose ecommerce-platform --local
 # This will automatically decompose ALL epics:
 #   - 01-infrastructure
 #   - 02-auth-backend
@@ -26,8 +26,22 @@ Break epic into concrete, actionable tasks.
 
 **Single Epic from Multi-Epic structure:**
 ```bash
-/pm:epic-decompose <feature_name>/<epic_folder>
-# Example: /pm:epic-decompose ecommerce-platform/01-infrastructure
+/pm:epic-decompose <feature_name>/<epic_folder> [--local]
+# Example: /pm:epic-decompose ecommerce-platform/01-infrastructure --local
+```
+
+## Flags
+
+`--local`, `-l`
+: Use local mode (offline workflow)
+: Creates task files in `.claude/epics/` directory structure
+: No GitHub/Azure synchronization
+: Task files remain local-only until manually synced
+: Ideal for offline work or projects without remote tracking
+
+Example:
+```
+/pm:epic-decompose user-authentication --local
 ```
 
 ## Required Documentation Access

package/autopm/.claude/commands/pm/prd-new.md

@@ -8,7 +8,20 @@ Launch interactive brainstorming session for new product requirement document.
 
 ## Usage
 ```
-/pm:prd-new <feature_name>
+/pm:prd-new <feature_name> [--local]
+```
+
+## Flags
+
+`--local`, `-l`
+: Use local mode (offline workflow)
+: Creates PRD files in `.claude/prds/` directory
+: No GitHub/Azure synchronization required
+: Ideal for working offline or without remote provider configured
+
+Example:
+```
+/pm:prd-new user-authentication --local
 ```
 
 ## Required Documentation Access

package/autopm/.claude/includes/task-creation-excellence.md

@@ -0,0 +1,18 @@
+## Task Creation Excellence
+
+When creating implementation tasks (for epic decomposition or project planning):
+
+1. **Query Context7 First**: Access latest agile/PM best practices
+2. **Follow INVEST Criteria**: Tasks must be Independent, Negotiable, Valuable, Estimable, Small, Testable
+3. **Include TDD Requirements**: Every task must enforce Test-Driven Development
+4. **Technology-Specific Details**: Leverage your technology-specific expertise in technical details
+5. **Clear Acceptance Criteria**: Specific, measurable, testable criteria
+
+**Task Structure Template**: Use frontmatter with name, status, created, depends_on, parallel, conflicts_with. Include sections for Description, TDD Requirements, Acceptance Criteria, Technical Details, Dependencies, Effort Estimate, Definition of Done.
+
+**Quality Standards**:
+- Tasks completable in 1-3 days maximum
+- Action-oriented titles (verb + noun)
+- Specific technical implementation guidance
+- Dependency and parallelization mapping
+- No vague descriptions or missing acceptance criteria

package/autopm/.claude/lib/ai-task-generator.js

@@ -0,0 +1,84 @@
+/**
+ * AI Task Generator
+ *
+ * Generates tasks from epic content using AI providers (OpenAI, Claude, etc.)
+ * Supports dependency injection for testing with mock providers.
+ *
+ * Usage:
+ *   const { TaskGenerator } = require('./ai-task-generator');
+ *
+ *   const generator = new TaskGenerator();
+ *   const tasks = await generator.generate(epicContent, { maxTasks: 10 });
+ */
+
+/**
+ * Task Generator using AI providers
+ */
+class TaskGenerator {
+  constructor(provider = null) {
+    this.provider = provider; // Allow injection for testing
+  }
+
+  /**
+   * Generate tasks from epic content
+   *
+   * @param {string} epicContent - Epic markdown content
+   * @param {Object} options - Generation options
+   * @param {number} [options.maxTasks=15] - Maximum number of tasks to generate
+   * @param {number} [options.minHours=2] - Minimum estimated hours per task
+   * @param {number} [options.maxHours=8] - Maximum estimated hours per task
+   * @returns {Promise<Array>} Array of task objects
+   */
+  async generate(epicContent, options = {}) {
+    const {
+      maxTasks = 15,
+      minHours = 2,
+      maxHours = 8
+    } = options;
+
+    if (!this.provider) {
+      throw new Error('AI provider not configured. Use TaskGenerator(provider) or set via setProvider');
+    }
+
+    // Generate tasks using AI provider
+    const tasks = await this.provider.generateTasks(epicContent, {
+      maxTasks,
+      minHours,
+      maxHours
+    });
+
+    // Validate and normalize tasks
+    return tasks.map((task, index) => this.normalizeTask(task, index + 1));
+  }
+
+  /**
+   * Normalize task object to standard format
+   *
+   * @param {Object} task - Raw task from AI
+   * @param {number} taskNumber - Task sequence number
+   * @returns {Object} Normalized task object
+   */
+  normalizeTask(task, taskNumber) {
+    return {
+      number: taskNumber,
+      title: task.title || `Task ${taskNumber}`,
+      description: task.description || '',
+      acceptance_criteria: task.acceptance_criteria || [],
+      estimated_hours: task.estimated_hours || 4,
+      dependencies: task.dependencies || [],
+      priority: task.priority || 'medium',
+      tags: task.tags || []
+    };
+  }
+
+  /**
+   * Set AI provider (for dependency injection)
+   *
+   * @param {Object} provider - AI provider instance
+   */
+  setProvider(provider) {
+    this.provider = provider;
+  }
+}
+
+module.exports = { TaskGenerator };

package/autopm/.claude/lib/cli-parser.js

@@ -0,0 +1,148 @@
+/**
+ * CLI Parser for PM Commands
+ *
+ * Provides unified command-line argument parsing with --local flag support
+ * across all PM commands using yargs best practices.
+ *
+ * Following Context7 yargs documentation patterns:
+ * - .boolean() for boolean flags
+ * - .alias() for short flags
+ * - .parse() for argument processing
+ */
+
+const yargs = require('yargs/yargs');
+const { hideBin } = require('yargs/helpers');
+const fs = require('fs');
+const path = require('path');
+
+/**
+ * Read project configuration to determine default provider
+ * @returns {string} Default provider ('github', 'azure', or 'local')
+ */
+function getDefaultProvider() {
+  try {
+    const configPath = path.join(process.cwd(), '.claude/config.json');
+    if (fs.existsSync(configPath)) {
+      const config = JSON.parse(fs.readFileSync(configPath, 'utf-8'));
+      return config.provider || 'github';
+    }
+  } catch (error) {
+    // Silently fall back to github if config cannot be read
+  }
+  return 'github';
+}
+
+/**
+ * Parse PM command arguments with --local flag support
+ *
+ * @param {string[]} args - Command arguments (usually process.argv)
+ * @returns {object} Parsed arguments with mode determined
+ *
+ * @example
+ * // Local mode
+ * parsePMCommand(['prd-new', 'feature', '--local'])
+ * // => { _: ['prd-new', 'feature'], local: true, mode: 'local' }
+ *
+ * @example
+ * // Remote mode (from config)
+ * parsePMCommand(['prd-list'])
+ * // => { _: ['prd-list'], local: false, mode: 'github' }
+ */
+function parsePMCommand(args) {
+  let failureMessage = null;
+
+  // Don't use hideBin if args are already clean (for testing)
+  const cleanArgs = Array.isArray(args) && !args[0]?.includes('node') ? args : hideBin(args);
+
+  const argv = yargs(cleanArgs)
+    .option('local', {
+      alias: 'l',
+      type: 'boolean',
+      describe: 'Use local mode (offline, no GitHub/Azure)',
+      default: false
+    })
+    .option('github', {
+      type: 'boolean',
+      describe: 'Use GitHub provider',
+      default: false
+    })
+    .option('azure', {
+      type: 'boolean',
+      describe: 'Use Azure DevOps provider',
+      default: false
+    })
+    .option('verbose', {
+      alias: 'v',
+      type: 'boolean',
+      describe: 'Enable verbose output',
+      default: false
+    })
+    .option('force', {
+      alias: 'f',
+      type: 'boolean',
+      describe: 'Force operation',
+      default: false
+    })
+    .option('output', {
+      alias: 'o',
+      type: 'string',
+      describe: 'Output format (json, text)',
+      choices: ['json', 'text']
+    })
+    .check((argv) => {
+      // Validate: cannot use --local with --github or --azure
+      if (argv.local && (argv.github || argv.azure)) {
+        throw new Error('Cannot use both --local and remote provider (--github or --azure)');
+      }
+      return true;
+    })
+    .fail((msg, err) => {
+      // Capture failure for throwing
+      failureMessage = msg || (err && err.message) || 'Unknown error';
+    })
+    .exitProcess(false)  // Don't call process.exit on error (for testing)
+    .parse();
+
+  // Throw error if validation failed
+  if (failureMessage) {
+    throw new Error(failureMessage);
+  }
+
+  // Determine mode based on flags and config
+  if (argv.local) {
+    argv.mode = 'local';
+  } else if (argv.github) {
+    argv.mode = 'github';
+  } else if (argv.azure) {
+    argv.mode = 'azure';
+  } else {
+    // Read from config
+    argv.mode = getDefaultProvider();
+  }
+
+  return argv;
+}
+
+/**
+ * Get help text for CLI parser
+ * Used for testing help text includes --local flag
+ *
+ * @returns {Promise<string>} Help text
+ */
+async function getHelpText() {
+  const parser = yargs([])
+    .option('local', {
+      alias: 'l',
+      type: 'boolean',
+      describe: 'Use local mode (offline, no GitHub/Azure)',
+      default: false
+    });
+
+  return await parser.getHelp();
+}
+
+module.exports = {
+  parsePMCommand,
+  getHelpText,
+  getDefaultProvider
+};