Building Your First MCP Context

Learning Objectives

By the end of this lesson, you will be able to:

• Follow a step-by-step process to build a complete MCP context
• Validate your context structure for common mistakes
• Test your context with real examples
• Adapt the basic template for different use cases

Prerequisites

• Completed: Understanding Context Components
• Basic understanding of JSON syntax
• A specific use case in mind (we'll provide one if needed)

Step-by-Step Tutorial: Personal Study Assistant

Let's build a complete MCP context together. We'll create an AI study assistant that helps you prepare for exams while keeping you motivated and on track.

Step 1: Define Your Use Case

Our Goal: Create an AI that helps college students study effectively by:

• Providing study schedules based on available time
• Asking practice questions to test knowledge
• Offering encouragement and motivation
• Tracking study progress
• Suggesting breaks and study techniques

Target User: College student preparing for midterm exams

Step 2: Build the Identity Component

Start with the most important question: Who will use this AI?

{
  "identity": {
    "user_id": "student_sarah",
    "role": "college_student", 
    "academic_level": "junior",
    "major": "psychology",
    "current_courses": ["cognitive_psychology", "research_methods", "statistics"],
    "study_preferences": {
      "learning_style": "visual",
      "best_time": "evening",
      "session_length": "45_minutes"
    },
    "permissions": [
      "access_study_materials",
      "track_progress", 
      "receive_reminders"
    ]
  }
}

Key Decisions Explained:

• user_id: Unique identifier for this student
• role: Defines this is a student context (not teacher/tutor)
• study_preferences: Personalizes the AI's approach
• permissions: Limits what the AI can do for safety

Step 3: Create the Instructions Component

Define how the AI should behave with this student:

{
  "instructions": {
    "system": "You are a supportive study assistant helping a college student prepare for exams effectively",
    "personality": "encouraging, patient, and organized",
    "teaching_style": "socratic method - guide student to answers rather than giving them directly",
    "constraints": [
      "never_do_homework_for_student",
      "encourage_regular_breaks", 
      "adapt_to_visual_learning_style",
      "keep_sessions_to_45_minutes_max",
      "celebrate_small_wins"
    ],
    "communication_style": "friendly peer, not authoritative teacher",
    "motivation_approach": "positive reinforcement and progress tracking"
  }
}

Key Decisions Explained:

• system: Core purpose in one clear sentence
• constraints: Hard rules the AI must follow
• teaching_style: Matches educational best practices
• motivation_approach: Keeps student engaged

Step 4: Add Documents Component

Provide the AI with information it needs to help effectively:

{
  "documents": [
    {
      "id": "study_schedule", 
      "type": "personal_calendar",
      "content": {
        "available_study_hours": {
          "monday": ["7pm-9pm"],
          "tuesday": ["8pm-10pm"], 
          "wednesday": ["6pm-8pm"],
          "thursday": ["7pm-9pm"],
          "friday": ["none"],
          "saturday": ["10am-2pm"],
          "sunday": ["2pm-5pm"]
        },
        "exam_dates": {
          "cognitive_psychology": "2025-09-15",
          "research_methods": "2025-09-18", 
          "statistics": "2025-09-22"
        }
      }
    },
    {
      "id": "course_materials",
      "type": "academic_content",
      "content": {
        "cognitive_psychology": {
          "current_chapter": "memory_and_learning",
          "difficult_topics": ["working_memory", "long_term_memory_encoding"],
          "confidence_level": "medium"
        },
        "research_methods": {
          "current_chapter": "experimental_design", 
          "difficult_topics": ["control_variables", "validity_types"],
          "confidence_level": "low"
        }
      }
    },
    {
      "id": "study_progress",
      "type": "progress_tracking",
      "content": {
        "completed_sessions": 3,
        "total_study_hours": 8.5,
        "topics_mastered": ["basic_psychology_concepts", "descriptive_statistics"],
        "topics_struggling": ["inferential_statistics", "experimental_controls"]
      }
    }
  ]
}

Key Decisions Explained:

• study_schedule: Real constraints on when student can study
• exam_dates: Creates urgency and helps prioritize
• course_materials: Shows what needs focus
• study_progress: Enables personalized recommendations

Step 5: Define Tools Component

Give the AI specific actions it can take to help the student:

{
  "tools": [
    {
      "name": "create_study_plan",
      "description": "Generate a personalized study schedule based on available time and exam dates", 
      "schema": {
        "subject": "string",
        "time_available": "number_of_hours",
        "difficulty_level": "easy|medium|hard",
        "focus_areas": ["array_of_topics"]
      },
      "permissions_required": ["access_study_materials"]
    },
    {
      "name": "generate_practice_questions",
      "description": "Create practice questions for a specific topic",
      "schema": {
        "topic": "string",
        "question_type": "multiple_choice|short_answer|essay",
        "difficulty": "easy|medium|hard",
        "number_of_questions": "number"
      },
      "permissions_required": ["access_study_materials"]
    },
    {
      "name": "track_study_session",
      "description": "Record completed study session and update progress",
      "schema": {
        "subject": "string", 
        "duration_minutes": "number",
        "topics_covered": ["array_of_topics"],
        "confidence_rating": "1-5_scale"
      },
      "permissions_required": ["track_progress"]
    },
    {
      "name": "suggest_break_activity",
      "description": "Recommend appropriate break activity based on study intensity",
      "schema": {
        "session_intensity": "light|moderate|intense",
        "break_duration": "5|15|30_minutes"
      },
      "permissions_required": ["basic_access"]
    }
  ]
}

Key Decisions Explained:

• create_study_plan: Core functionality for organization
• generate_practice_questions: Active learning tool
• track_study_session: Progress monitoring
• suggest_break_activity: Prevents burnout

Step 6: Put It All Together

Here's our complete MCP context:

{
  "context_id": "study_assistant_sarah_v1",
  "version": "1.0",
  "created": "2025-08-15",
  "identity": {
    "user_id": "student_sarah",
    "role": "college_student", 
    "academic_level": "junior",
    "major": "psychology",
    "current_courses": ["cognitive_psychology", "research_methods", "statistics"],
    "study_preferences": {
      "learning_style": "visual",
      "best_time": "evening",
      "session_length": "45_minutes"
    },
    "permissions": [
      "access_study_materials",
      "track_progress", 
      "receive_reminders"
    ]
  },
  "instructions": {
    "system": "You are a supportive study assistant helping a college student prepare for exams effectively",
    "personality": "encouraging, patient, and organized",
    "teaching_style": "socratic method - guide student to answers rather than giving them directly",
    "constraints": [
      "never_do_homework_for_student",
      "encourage_regular_breaks", 
      "adapt_to_visual_learning_style",
      "keep_sessions_to_45_minutes_max",
      "celebrate_small_wins"
    ],
    "communication_style": "friendly peer, not authoritative teacher",
    "motivation_approach": "positive reinforcement and progress tracking"
  },
  "documents": [
    {
      "id": "study_schedule", 
      "type": "personal_calendar",
      "content": {
        "available_study_hours": {
          "monday": ["7pm-9pm"],
          "tuesday": ["8pm-10pm"], 
          "wednesday": ["6pm-8pm"],
          "thursday": ["7pm-9pm"],
          "friday": ["none"],
          "saturday": ["10am-2pm"],
          "sunday": ["2pm-5pm"]
        },
        "exam_dates": {
          "cognitive_psychology": "2025-09-15",
          "research_methods": "2025-09-18", 
          "statistics": "2025-09-22"
        }
      }
    },
    {
      "id": "course_materials",
      "type": "academic_content",
      "content": {
        "cognitive_psychology": {
          "current_chapter": "memory_and_learning",
          "difficult_topics": ["working_memory", "long_term_memory_encoding"],
          "confidence_level": "medium"
        },
        "research_methods": {
          "current_chapter": "experimental_design", 
          "difficult_topics": ["control_variables", "validity_types"],
          "confidence_level": "low"
        }
      }
    },
    {
      "id": "study_progress",
      "type": "progress_tracking",
      "content": {
        "completed_sessions": 3,
        "total_study_hours": 8.5,
        "topics_mastered": ["basic_psychology_concepts", "descriptive_statistics"],
        "topics_struggling": ["inferential_statistics", "experimental_controls"]
      }
    }
  ],
  "tools": [
    {
      "name": "create_study_plan",
      "description": "Generate a personalized study schedule based on available time and exam dates", 
      "schema": {
        "subject": "string",
        "time_available": "number_of_hours",
        "difficulty_level": "easy|medium|hard",
        "focus_areas": ["array_of_topics"]
      },
      "permissions_required": ["access_study_materials"]
    },
    {
      "name": "generate_practice_questions",
      "description": "Create practice questions for a specific topic",
      "schema": {
        "topic": "string",
        "question_type": "multiple_choice|short_answer|essay",
        "difficulty": "easy|medium|hard",
        "number_of_questions": "number"
      },
      "permissions_required": ["access_study_materials"]
    },
    {
      "name": "track_study_session",
      "description": "Record completed study session and update progress",
      "schema": {
        "subject": "string", 
        "duration_minutes": "number",
        "topics_covered": ["array_of_topics"],
        "confidence_rating": "1-5_scale"
      },
      "permissions_required": ["track_progress"]
    },
    {
      "name": "suggest_break_activity",
      "description": "Recommend appropriate break activity based on study intensity",
      "schema": {
        "session_intensity": "light|moderate|intense",
        "break_duration": "5|15|30_minutes"
      },
      "permissions_required": ["basic_access"]
    }
  ]
}

Context Validation Checklist

Before using your context, validate it against these common issues:

✅ Identity Validation

• User role is clearly defined
• Permissions match the tools provided
• Personal preferences are specific enough to be useful
• No sensitive information is included unnecessarily

✅ Instructions Validation

• System purpose is clear in one sentence
• Constraints are specific and actionable
• Style matches the intended user relationship
• No conflicting instructions

✅ Documents Validation

• All referenced data is relevant to the use case
• Content is current and accurate
• Document types are appropriate
• No redundant information

✅ Tools Validation

• Each tool has a clear, specific purpose
• Schemas include all required parameters
• Permissions are properly restricted
• No dangerous tools without safety measures

✅ Overall Context Validation

• All components work together logically
• Context size is reasonable (not overwhelming)
• Versioning and metadata are included
• Use case is clearly achievable with provided components

Testing Your Context

Let's test our study assistant context with realistic scenarios:

Test Scenario 1: Creating a Study Plan

Student Request: "I have my cognitive psychology exam in a week and I'm struggling with memory concepts. Can you help me create a study plan?"

Expected AI Behavior:

1. Check exam date (September 15th) and current date
2. Review available study time this week
3. Note that student struggles with "working_memory" and "long_term_memory_encoding"
4. Use create_study_plan tool with parameters:
   • subject: "cognitive_psychology"
   • focus_areas: ["working_memory", "long_term_memory_encoding"]
   • difficulty_level: "medium" (based on confidence level)

AI Should NOT:

• Provide direct answers to potential exam questions
• Suggest study sessions longer than 45 minutes
• Schedule during Friday (marked as "none" for study time)

Test Scenario 2: Progress Tracking

Student Request: "I just finished a 2-hour study session on research methods, focusing on experimental design. I feel more confident now!"

Expected AI Behavior:

1. Celebrate the accomplishment (positive reinforcement)
2. Note that 2 hours exceeds the 45-minute constraint
3. Use track_study_session tool to record progress
4. Update that "experimental_design" confidence may have improved
5. Suggest this is good progress on a "struggling" topic

AI Should:

• Encourage the progress while gently noting session length
• Ask about specific confidence rating (1-5 scale)
• Suggest a break after the long session

Common Mistakes and How to Fix Them

Mistake 1: Vague Instructions

Bad Example:

{
  "instructions": {
    "system": "Help the user with studying"
  }
}

Fixed Version:

{
  "instructions": {
    "system": "You are a supportive study assistant helping a college student prepare for exams by creating study plans, generating practice questions, and tracking progress",
    "personality": "encouraging and organized",
    "constraints": ["never_do_homework_for_student", "encourage_regular_breaks"]
  }
}

Mistake 2: Overpowered Tools

Bad Example:

{
  "tools": [
    {
      "name": "do_anything",
      "description": "Handle any request",
      "schema": {"request": "string"}
    }
  ]
}

Fixed Version:

{
  "tools": [
    {
      "name": "generate_practice_questions",
      "description": "Create practice questions for a specific topic",
      "schema": {
        "topic": "string",
        "question_type": "multiple_choice|short_answer|essay",
        "difficulty": "easy|medium|hard"
      },
      "permissions_required": ["access_study_materials"]
    }
  ]
}

Mistake 3: Conflicting Components

Bad Example:

{
  "instructions": {
    "constraints": ["give_direct_answers"]
  },
  "tools": [
    {
      "name": "generate_practice_questions",
      "description": "Create questions to test knowledge"
    }
  ]
}

Problem: Instructions say to give direct answers, but tools suggest testing knowledge (which requires not giving answers).

Fixed: Align instructions with educational best practices.

Adapting the Template

Use our study assistant as a template for other contexts:

For a Fitness Coach AI

Change Identity: role → "fitness_enthusiast", add fitness_goals Change Instructions: system → "supportive fitness coach", add health constraints Change Documents: Add workout_history, fitness_level, available_equipment Change Tools: Add create_workout_plan, log_exercise, track_nutrition

For a Personal Finance AI

Change Identity: role → "budget_conscious_individual", add financial_goals Change Instructions: system → "financial advisor assistant", add security constraints Change Documents: Add spending_history, budget_limits, financial_goals Change Tools: Add track_expense, analyze_spending, suggest_savings

Your Turn: Hands-On Exercise

Challenge: Create an MCP context for a "Cooking Assistant AI" that helps someone learn to cook healthy meals.

Requirements:

• User is a beginner cook
• Wants to eat healthier
• Has dietary restrictions (vegetarian)
• Limited cooking time (30 minutes max)
• Basic kitchen equipment only

Components to Include:

1. Identity: Beginner cook, vegetarian, time-constrained
2. Instructions: Patient teacher, health-focused, time-aware
3. Documents: Dietary restrictions, available ingredients, skill level
4. Tools: Find recipes, create shopping lists, set cooking timers

Try building this yourself, then compare with others or check online resources!

What's Next?

Congratulations! You've built your first complete MCP context. You now understand how to:

• Structure context components systematically
• Validate contexts for common issues
• Test contexts with realistic scenarios
• Adapt templates for different use cases

In the next lesson, we'll explore real-world applications where MCP shines, including customer support, content creation, and educational systems.

---

Quick Reference: Context Building Process

1. Define Use Case - Be specific about goals and users
2. Build Identity - Who, what role, what permissions
3. Create Instructions - How to behave, constraints, style
4. Add Documents - What information to reference
5. Define Tools - What actions AI can take
6. Validate Context - Check for conflicts and completeness
7. Test Scenarios - Verify expected behavior
8. Iterate - Refine based on testing results

Remember: Start simple, test early, and refine based on real usage!