Agent Knowledge MCP

#!/usr/bin/env python3 """ Test the complete enhanced smart prompting system including AI guidance with citations. """ import asyncio import tempfile import sys from pathlib import Path # Add src to path sys.path.append(str(Path(__file__).parent.parent)) from fastmcp import Client from src.prompts.sub_servers.smart_prompting_server import app as smart_prompting_app async def test_complete_enhanced_system(): """Test the complete enhanced smart prompting system with AI citations.""" print("🚀 Testing Complete Enhanced Smart Prompting System") print("=" * 60) # Create temporary .knowledges structure in project root project_root = Path(__file__).parent.parent knowledges_dir = project_root / ".knowledges_test" try: # Create test structure workflows_dir = knowledges_dir / "workflows" rules_dir = knowledges_dir / "rules" memories_dir = knowledges_dir / "memories" workflows_dir.mkdir(parents=True, exist_ok=True) rules_dir.mkdir(parents=True, exist_ok=True) memories_dir.mkdir(parents=True, exist_ok=True) # Create sample content workflow_content = """# API Development Workflow ## Development Steps 1. Design API endpoints 2. Write OpenAPI specification 3. Implement handlers 4. Add validation middleware 5. Write comprehensive tests 6. Update documentation ## Code Review Process 1. Self-review checklist 2. Peer review assignment 3. Address feedback 4. Final approval and merge ## Deployment 1. Run CI/CD pipeline 2. Deploy to staging environment 3. Run integration tests 4. Deploy to production""" rules_content = """# FastMCP Development Rules ## Architecture Standards - Use FastMCP framework for all MCP servers - Implement proper error handling - Add comprehensive logging - Follow async/await patterns ## Code Quality - Type hints are mandatory - Docstrings for all public functions - Unit tests with >80% coverage - Use proper exception handling ## Security Guidelines - Validate all user inputs - Use secure authentication - Log security events - Regular dependency updates""" memory_content = """# Project Lessons Learned ## FastMCP Integration Issues Date: 2025-01-15 Issue: Client testing required specific attribute handling Solution: Use result.data for tools, result.messages[0].content.text for prompts Location: tests/test_modular_prompt_server.py:lines 45-50 ## Virtual Environment Setup Date: 2025-01-20 Issue: macOS externally-managed-environment restrictions Solution: Always use virtual environments for testing Command: python3 -m venv test_env && source test_env/bin/activate ## Sub Server Architecture Date: 2025-01-25 Lesson: Modular mounting pattern works well with FastMCP Benefits: 64% code reduction while maintaining functionality Files: src/prompts/prompt_server.py, src/prompts/sub_servers/""" # Write files (workflows_dir / "api-development.md").write_text(workflow_content) (rules_dir / "fastmcp-standards.md").write_text(rules_content) (memories_dir / "project-lessons.md").write_text(memory_content) print(f"📁 Created test knowledge base: {knowledges_dir}") # Test with FastMCP client (simulating no workspace - will show formatted error) print(f"\n🧪 Testing smart prompting tool with enhanced citations...") async with Client(smart_prompting_app) as client: # List available tools tools = await client.list_tools() print(f"🛠️ Available tools: {[tool.name for tool in tools]}") # Test the ask_mcp_advance tool try: result = await client.call_tool("ask_mcp_advance", { "intended_action": "implement new API endpoint", "task_description": "Create a new REST API endpoint for user management with proper validation and testing", "scope": "feature" }) guidance = result.data print(f"\n📋 **Enhanced Guidance Response:**") print("-" * 50) print(guidance) print("-" * 50) # Verify enhancement features in response print(f"\n✅ **Enhancement Verification:**") # Check that the response asks for citations (even though workspace isn't available) has_citation_requirement = "file citations" in guidance.lower() or "backticks" in guidance.lower() print(f" 📝 Citation requirement mentioned: {'✅ YES' if has_citation_requirement else '❌ NO'}") # Check error handling for no workspace has_workspace_guidance = "workspace" in guidance.lower() and "permissions" in guidance.lower() print(f" 🏠 Workspace guidance provided: {'✅ YES' if has_workspace_guidance else '❌ NO'}") # Verify structured format has_structured_format = "**" in guidance and "Error" in guidance print(f" 📋 Structured format: {'✅ YES' if has_structured_format else '❌ NO'}") except Exception as e: print(f"❌ Tool call failed: {str(e)}") print(f"\n🎯 **Expected Behavior with Real Workspace:**") print(f" 📁 Tool would load from: `{knowledges_dir}/`") print(f" 📄 Files with line numbers: workflows/api-development.md:1-20") print(f" 🤖 AI would include citations like: `rules/fastmcp-standards.md:8-12`") print(f" 📝 User could then ask agent to edit specific lines") print(f"\n💡 **Citation Examples AI Should Provide:**") print(f" ✅ 'Follow API development process in `workflows/api-development.md:3-8`'") print(f" ✅ 'Apply FastMCP standards from `rules/fastmcp-standards.md:4-7`'") print(f" ✅ 'Remember testing lessons from `memories/project-lessons.md:3-5`'") return True finally: # Cleanup test directory if knowledges_dir.exists(): import shutil shutil.rmtree(knowledges_dir) print(f"\n🧹 Cleaned up test directory: {knowledges_dir}") if __name__ == "__main__": asyncio.run(test_complete_enhanced_system())