VasperaPM

╔═══════════════════════════════════════════════════════════════╗
║                                                               ║
║   ██╗   ██╗ █████╗ ███████╗██████╗ ███████╗██████╗  █████╗    ║
║   ██║   ██║██╔══██╗██╔════╝██╔══██╗██╔════╝██╔══██╗██╔══██╗   ║
║   ██║   ██║███████║███████╗██████╔╝█████╗  ██████╔╝███████║   ║
║   ╚██╗ ██╔╝██╔══██║╚════██║██╔═══╝ ██╔══╝  ██╔══██╗██╔══██║   ║
║    ╚████╔╝ ██║  ██║███████║██║     ███████╗██║  ██║██║  ██║   ║
║     ╚═══╝  ╚═╝  ╚═╝╚══════╝╚═╝     ╚══════╝╚═╝  ╚═╝╚═╝  ╚═╝   ║
║                                                               ║
║              AI-Powered Product Management                    ║
║                                                               ║
╚═══════════════════════════════════════════════════════════════╝

Transform your ideas into production-ready specs with AI

What is VasperaPM?

VasperaPM is an MCP (Model Context Protocol) server that brings AI-powered product management tools directly into Claude Code. Generate PRDs, API specs, test cases, and sync with your favorite PM tools - all from your IDE.

Key Features

• Generate PRDs - Create comprehensive Product Requirements Documents from context
• Infer from Code - Analyze existing codebases to generate documentation
• Verify Code - Detect when code drifts from requirements
• API Specs - Auto-generate OpenAPI/Swagger specifications
• Test Cases - Create detailed test plans and acceptance criteria
• PM Tool Sync - Export to Jira, Linear, GitHub, and Azure DevOps

What's New in v2.10

All 18 core tools now at 5-star quality with:

• Error flow analysis and quality assessments
• Context-aware scoring and evidence-based confidence
• Industry-specific compliance checks
• Interactive setup validation and runnable checklists
• Three-tier severity classification for PRD reviews

Quick Start

# One command - that's it
npx vaspera-pm

# Then reload VS Code (Cmd+Shift+P → "Developer: Reload Window")
# Ask Claude: "Use vaspera-pm to analyze this codebase"

This single command:

• ✅ Finds or creates your API key
• ✅ Validates your credentials
• ✅ Configures .mcp.json for Claude Code
• ✅ Works with VasperaPM cloud OR your own Anthropic key

Run it again anytime to update or fix your setup.

Commands

| Command | Description | |---------|-------------| | vaspera-pm connect | Link to VasperaPM account (browser OAuth flow) | | vaspera-pm install | Configure MCP server for Claude Code | | vaspera-pm analyze <path> | Full analysis: discover → analyze → verify → output | | vaspera-pm verify <path> | Quick verification check (for CI/CD) | | vaspera-pm fix <path> | Generate and apply fix suggestions for drift findings | | vaspera-pm discover <path> | Discovery only (no AI, no API key needed) | | vaspera-pm init | Initialize VasperaPM in a project | | vaspera-pm serve | Start the MCP server (used by Claude Code) | | vaspera-pm status | Check your current configuration |

MCP 2025 Features

VasperaPM supports the latest MCP 2025 specification features:

Tool Annotations

All tools include behavioral metadata hints for better client UX:

• readOnlyHint - Indicates if tool modifies state
• destructiveHint - Warns about destructive operations
• idempotentHint - Whether repeated calls are safe
• openWorldHint - If tool accesses external systems

Structured Outputs

Key tools return structuredContent alongside markdown for programmatic use:

• validate_implementation - Coverage reports with requirements status
• explode_backlog - Array of user stories with estimates
• dependency_audit - Vulnerability findings

Async Tasks (Experimental)

Long-running tools support async execution with progress tracking:

# Request async execution with _task: true
# Tool returns task ID immediately
# Use tasks_status, tasks_result, tasks_list, tasks_cancel to manage

# Async-capable tools:
# - run_pipeline
# - verify_docs_code
# - generate_complete_backlog

Available Tools (30+ Total)

Once installed, these tools are available in Claude Code:

Requirements & PRDs

| Tool | Description | |------|-------------| | synthesize_requirements | Extract requirements from meeting notes, emails, specs | | synthesize_master_prd | Create unified PRD from multiple sources | | infer_prd_from_code | Reverse-engineer PRD from existing code | | review_prd | Quality review with completeness scoring | | explode_backlog | Break features into sprint-ready stories |

Code Analysis

| Tool | Description | |------|-------------| | explain_codebase | Generate onboarding documentation | | reverse_engineer_user_flows | Extract user journeys from code | | validate_implementation | Verify code matches requirements | | suggest_refactors | Identify tech debt and improvements |

Architecture & Specs

| Tool | Description | |------|-------------| | generate_architecture | Create system design from requirements | | generate_api_docs | Auto-generate OpenAPI specifications | | generate_test_specs | Create comprehensive test plans | | handoff_package | Generate developer handoff documentation |

Verification & Drift Detection

| Tool | Description | |------|-------------| | verify_docs_code | Detect documentation vs code drift | | detect_contradictions | Find conflicting requirements across sources | | analyze_git_freshness | Find stale documentation | | dependency_audit | Security and update checks |

Export & Integration

| Tool | Description | |------|-------------| | sync_to_tracker | Export to Jira, Linear, GitHub, ADO | | export_requirements | Generate portable requirement formats | | generate_complete_backlog | Full sprint backlog generation |

DevOps & Traceability

| Tool | Description | |------|-------------| | generate_deployment_runbook | Create deployment documentation | | generate_changelog | Auto-generate release notes | | detect_regressions | Find breaking changes | | generate_sprint_report | Sprint summary and metrics |

Pipeline Automation

| Tool | Description | |------|-------------| | run_pipeline | Orchestrate multi-tool workflows |

Task Management (MCP 2025)

| Tool | Description | |------|-------------| | tasks_status | Get status of a running task | | tasks_result | Get result of a completed task | | tasks_list | List all tasks with filtering | | tasks_cancel | Cancel a pending or running task |

API Keys

Option 1: Local Mode (Your Own Key)

Use your own Anthropic API key - no account needed:

export ANTHROPIC_API_KEY=sk-ant-xxx
npx vaspera-pm install
# or
npx vaspera-pm analyze .

Features:

• All 30 tools work locally
• Output saved to docs/vasperaPM/
• No account required
• Works in air-gapped environments
• Unlimited usage (your API costs)

Option 2: VasperaPM Cloud

Connect to VasperaPM for managed usage and team features:

npx vaspera-pm connect  # Opens browser for login
npx vaspera-pm install  # Configures Claude Code

Features:

• Team collaboration & sharing
• Usage tracking & quotas (Free: 100, Starter: 500, Pro: 2,500/month)
• Jira/Linear/GitHub/ADO integrations
• Historical drift tracking
• No Anthropic key needed (we proxy requests)

Configuration

VasperaPM uses sensible defaults - no configuration required.

Configuration locations:

• .mcp.json - Project-level MCP config (Claude Code reads this)
• .vaspera/config.json - Project credentials (created by connect)
• ~/.vasperapm/config.json - Global settings
• VASPERA_API_KEY - Environment variable for cloud mode
• ANTHROPIC_API_KEY - Environment variable for local mode

Integrations

Connect your PM tools at vasperapm.com/dashboard/integrations:

• Jira - Sync PRDs to epics and stories
• Linear - Export tasks to Linear projects
• GitHub - Create issues and track progress

Requirements

• Node.js 20+
• Claude Code (VSCode extension or CLI)

Support

• Documentation: github.com/rcolkitt/VasperaPM
• Issues: github.com/rcolkitt/VasperaPM/issues
• Dashboard: vasperapm.com/dashboard

License

Proprietary License - see LICENSE for details.