MCP Ollama Consult Server

# MCP Ollama Consult - Usage Examples Comprehensive examples demonstrating how to use MCP Ollama Consult for various scenarios. ## Table of Contents - [Basic Consultation](#basic-consultation) - [Model Comparison](#model-comparison) - [Sequential Reasoning Chains](#sequential-reasoning-chains) - [Memory Management](#memory-management) - [Real-World Scenarios](#real-world-scenarios) - [Integration Patterns](#integration-patterns) - [Advanced Usage](#advanced-usage) ## Basic Consultation ### Simple Question & Answer ```typescript { tool: "consult_ollama", arguments: { prompt: "What is the difference between REST and GraphQL?", model: "llama3.2" } } ``` ### Code Review ```typescript { tool: "consult_ollama", arguments: { prompt: `Review this React component for best practices: function UserProfile({ userId }) { const [user, setUser] = useState(null); useEffect(() => { fetch('/api/users/' + userId) .then(response => response.json()) .then(setUser); }, [userId]); return <div>{user?.name}</div>; }`, model: "qwen2.5-coder:7b", context: "This is for a production React application with TypeScript", systemPrompt: "You are a senior React developer. Focus on performance, security, and maintainability." } } ``` ### Architecture Consultation ```typescript { tool: "consult_ollama", arguments: { prompt: "Design a scalable architecture for a real-time chat application that needs to support 100K concurrent users", model: "deepseek-v3.1", context: "Budget constraints require cloud-native solutions, team has experience with Node.js and React", timeoutMs: 300000 } } ``` ## Model Comparison ### Technology Decision ```typescript { tool: "compare_ollama_models", arguments: { prompt: "Should we use Kubernetes or Docker Swarm for our microservices deployment?", models: ["llama3.2", "qwen2.5-coder:7b", "deepseek-v3.1"], context: "Medium-sized startup, 15-person engineering team, budget-conscious, needs to deploy 8 microservices" } } ``` ### Code Solution Alternatives ```typescript { tool: "compare_ollama_models", arguments: { prompt: "Implement a rate limiter in Python with these requirements: 100 requests per minute per user, memory-efficient, thread-safe", models: ["qwen2.5-coder:7b", "deepseek-v3.1"], context: "Flask web application, Redis available, performance is critical", timeoutMs: 180000 } } ``` ### Design Pattern Evaluation ```typescript { tool: "compare_ollama_models", arguments: { prompt: "Compare Observer vs Pub/Sub patterns for this use case: user activity tracking in a social media app", models: ["llama3.2", "qwen2.5-coder:7b"], context: "Need to track likes, comments, shares, follows across multiple services" } } ``` ## Sequential Reasoning Chains ### Code Analysis to Implementation ```typescript { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "analyzer", model: "llama3.2", prompt: `Analyze this legacy PHP code for refactoring opportunities: function processOrder($order) { global $db, $email, $inventory; if (!$order) return false; if ($order['total'] > 1000) { $discount = 0.1; $order['total'] *= (1 - $discount); } $query = "INSERT INTO orders VALUES ('" . $order['id'] . "', '" . $order['total'] . "')"; mysql_query($query); if ($order['total'] > 500) { mail($order['email'], 'Order Confirmation', 'Your order has been processed'); } return true; }`, timeoutMs: 120000 }, { id: "architect", model: "qwen2.5-coder:7b", prompt: "Based on {analyzer}'s analysis, design a modern object-oriented architecture that addresses the identified issues", timeoutMs: 180000 }, { id: "implementer", model: "deepseek-v3.1", prompt: "Using {architect}'s design, write clean, modern PHP code that implements the proposed architecture", timeoutMs: 240000 } ] } } ``` ### Security Review Process ```typescript { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "security_auditor", model: "llama3.2", prompt: `Perform security audit of this authentication endpoint: POST /api/login { "username": "user@example.com", "password": "plaintext_password" } Current implementation: - Stores passwords in MD5 - No rate limiting - JWT tokens never expire - No HTTPS enforcement`, systemPrompt: "You are a cybersecurity expert specializing in web application security", timeoutMs: 150000 }, { id: "security_architect", model: "qwen2.5-coder:7b", prompt: "Based on {security_auditor}'s findings, design a comprehensive security strategy with specific technical recommendations", timeoutMs: 180000 }, { id: "developer", model: "deepseek-v3.1", prompt: "Implement {security_architect}'s recommendations with working code examples and configuration", timeoutMs: 300000 } ] } } ``` ### Product Planning Chain ```typescript { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "market_analyst", model: "llama3.2", prompt: "Analyze market opportunity for a developer productivity tool that helps with code reviews using AI", context: "Target audience: mid-size tech companies (50-500 developers)" }, { id: "product_manager", model: "qwen2.5-coder:7b", prompt: "Based on {market_analyst}'s research, define MVP features and roadmap for the code review AI tool" }, { id: "tech_lead", model: "deepseek-v3.1", prompt: "Using {product_manager}'s requirements, design the technical architecture and implementation plan" } ] } } ``` ## Memory Management ### Storing Analysis Results ```typescript // First, get architectural guidance { tool: "consult_ollama", arguments: { prompt: "Design microservices architecture for e-commerce platform with these requirements: user management, product catalog, order processing, payment integration, inventory management", model: "deepseek-v3.1", context: "Expected 50K daily active users, budget $5K/month for infrastructure", timeoutMs: 300000 } } // Then store the result for future reference { tool: "remember_consult", arguments: { key: "ecommerce_architecture_2024_v1", value: "< response from above consultation >", metadata: { model: "deepseek-v3.1", topic: "microservices_architecture", project: "ecommerce_platform", requirements: "50k_dau_5k_budget", version: "1.0" }, tags: ["architecture", "microservices", "ecommerce", "scalability"] } } ``` ### Building Knowledge Base ```typescript // Store coding standards { tool: "remember_consult", arguments: { key: "typescript_coding_standards", value: "Our TypeScript coding standards: 1. Use strict mode, 2. Prefer interfaces over types for object shapes, 3. Use enum for constants...", tags: ["coding-standards", "typescript", "team-guidelines"] } } // Store architectural decisions { tool: "remember_consult", arguments: { key: "adr_001_database_choice", value: "ADR-001: Use PostgreSQL for main database. Reasoning: ACID compliance needed, team expertise, excellent JSON support...", metadata: { type: "architectural_decision", status: "accepted", date: "2024-11-17" }, tags: ["adr", "database", "postgresql"] } } ``` ## Real-World Scenarios ### Debugging Session ```typescript { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "debugger", model: "qwen2.5-coder:7b", prompt: `Help debug this performance issue: Symptom: Web page takes 8-12 seconds to load Stack: React frontend, Node.js/Express backend, PostgreSQL Database queries are fast (< 100ms) Network requests show 200ms RTT Chrome DevTools shows: - Bundle size: 2.3MB - 47 JavaScript files loaded - 23 CSS files loaded - Total requests: 156 - Largest contentful paint: 8.2s`, systemPrompt: "You are a performance optimization specialist" }, { id: "optimizer", model: "deepseek-v3.1", prompt: "Based on {debugger}'s analysis, provide step-by-step optimization plan with specific webpack, React, and bundling improvements", timeoutMs: 240000 } ] } } ``` ### Code Refactoring Consultation ```typescript { tool: "compare_ollama_models", arguments: { prompt: `Refactor this legacy jQuery code to modern vanilla JavaScript: $('#user-form').on('submit', function(e) { e.preventDefault(); var formData = $(this).serialize(); $.ajax({ url: '/api/users', type: 'POST', data: formData, success: function(response) { $('#message').html('<div class="success">User created!</div>'); $('#user-form')[0].reset(); }, error: function(xhr, status, error) { $('#message').html('<div class="error">Error: ' + error + '</div>'); } }); });`, models: ["qwen2.5-coder:7b", "deepseek-v3.1"], context: "Must support IE11, no external dependencies allowed, prefer modern ES6+ features where possible" } } ``` ### API Design Review ```typescript { tool: "consult_ollama", arguments: { prompt: `Review this REST API design for a blog platform: GET /posts - list all posts GET /posts/:id - get specific post POST /posts - create post (requires auth) PUT /posts/:id - update post (requires auth + ownership) DELETE /posts/:id - delete post (requires auth + ownership) GET /users/:id/posts - get user's posts GET /posts/:id/comments - get post comments POST /posts/:id/comments - add comment (requires auth) POST /auth/login - login POST /auth/register - register POST /auth/logout - logout What improvements would you suggest?`, model: "llama3.2", context: "Planning to support mobile app, expecting 10K users, using JWT tokens", systemPrompt: "You are a senior backend architect with expertise in REST API design" } } ``` ## Integration Patterns ### With CI/CD Pipeline ```typescript // Code quality consultation in CI { tool: "consult_ollama", arguments: { prompt: `Analyze this pull request for code quality issues: Files changed: 3 Lines added: 247 Lines removed: 89 Key changes: - Refactored user authentication service - Added password complexity validation - Updated API error handling - Added unit tests for new validation logic Code complexity increased from 6.2 to 8.7 Test coverage decreased from 89% to 84% Should this PR be approved?`, model: "qwen2.5-coder:7b", context: "Team follows strict code quality standards, deadline pressure is moderate" } } ``` ### With Development Tools ```typescript // Architecture decision support { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "requirements_analyst", model: "llama3.2", prompt: "Analyze requirements for choosing between GraphQL and REST for this API: social media platform with user profiles, posts, comments, real-time notifications" }, { id: "architect", model: "qwen2.5-coder:7b", prompt: "Based on {requirements_analyst}, make specific technology recommendation with pros/cons" } ] } } ``` ### With Monitoring/Alerts ```typescript // Performance incident analysis { tool: "consult_ollama", arguments: { prompt: `Analyze this performance incident: Time: 2024-11-17 14:30-15:45 UTC Duration: 75 minutes Impact: 40% increase in response time Metrics during incident: - CPU usage: 85% (normal: 45%) - Memory usage: 92% (normal: 60%) - Database connections: 95/100 (normal: 40/100) - Queue depth: 2,847 jobs (normal: 12 jobs) Recent deployments: - User service v2.3.1 (deployed 1 hour before incident) - Updated database schema (3 hours before) What likely caused this and how to prevent it?`, model: "deepseek-v3.1", context: "Node.js microservices, PostgreSQL database, Redis queue, AWS infrastructure", timeoutMs: 180000 } } ``` ## Advanced Usage ### Custom System Prompts ```typescript { tool: "consult_ollama", arguments: { prompt: "How should we structure our React components for this e-commerce site?", model: "qwen2.5-coder:7b", systemPrompt: `You are a React expert with these constraints: - Must use function components only - TypeScript strict mode required - Performance is critical (mobile users on slow connections) - Team prefers composition over inheritance - Must be accessible (WCAG 2.1 AA compliance) - Prefer CSS modules over styled-components` } } ``` ### Context-Rich Consultations ```typescript { tool: "consult_ollama", arguments: { prompt: "Should we migrate this service to Kubernetes?", model: "llama3.2", context: `Current setup: - Single VM running Docker Compose - 4 services: web, api, database, redis - 500 daily active users - 99.2% uptime over 6 months - Team: 3 developers, 1 devops engineer - Budget: $800/month for infrastructure - Growth: 20% monthly user increase - Pain points: deployment downtime, scaling web service manually Team experience: - Strong Docker knowledge - Basic AWS experience - No Kubernetes experience - Limited time for learning (startup phase)` } } ``` ### Chain with Error Handling ```typescript { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "reviewer", model: "llama3.2", prompt: "Review this database schema for a booking system", timeoutMs: 90000 }, { id: "optimizer", model: "qwen2.5-coder:7b", prompt: "If {reviewer} found issues, provide optimization suggestions. If no issues, suggest performance enhancements.", timeoutMs: 120000 }, { id: "implementer", model: "deepseek-v3.1", prompt: "Based on {optimizer}'s recommendations, provide SQL migration scripts and updated schema", timeoutMs: 180000 } ] } } ``` ### Multi-Language Consultation ```typescript { tool: "compare_ollama_models", arguments: { prompt: "Implement bubble sort algorithm with detailed comments explaining each step", models: ["qwen2.5-coder:7b", "deepseek-v3.1"], context: "Need implementations in Python, JavaScript, and Go. Focus on readability for teaching purposes." } } ``` ## Best Practices ### Timeout Management ```typescript // Short timeout for quick questions { tool: "consult_ollama", arguments: { prompt: "What HTTP status code should I use for 'resource not found'?", model: "llama3.2", timeoutMs: 30000 } } // Long timeout for complex analysis { tool: "consult_ollama", arguments: { prompt: "Perform comprehensive security audit of this authentication system...", model: "deepseek-v3.1", timeoutMs: 600000 // 10 minutes } } ``` ### Memory Organization ```typescript // Use structured keys { tool: "remember_consult", arguments: { key: "project_alpha/architecture/auth_service/v2.1", value: "Authentication service architecture...", tags: ["project-alpha", "auth", "architecture", "v2.1"] } } // Include search metadata { tool: "remember_consult", arguments: { key: "team_decisions/database_choice_2024", value: "Decision: PostgreSQL for primary database...", metadata: { decision_date: "2024-11-17", participants: ["john", "sarah", "mike"], alternatives_considered: ["MySQL", "MongoDB"], review_date: "2025-05-17" } } } ``` ### Chain Optimization ```typescript // Use specific, focused consultants { tool: "sequential_consultation_chain", arguments: { consultants: [ { id: "security_scanner", model: "llama3.2", prompt: "Scan for security vulnerabilities only", timeoutMs: 60000 }, { id: "security_fixer", model: "qwen2.5-coder:7b", prompt: "Fix vulnerabilities found by {security_scanner}", timeoutMs: 120000 } ] } } ```