🎯 Prompt Challenge Mode

Test your prompt engineering skills with these progressively harder challenges. Each challenge gives you a goal — write the best prompt to achieve it!

---

🟢 Beginner Challenges

Challenge 1: The Specific Summarizer

Goal: Write a prompt that summarizes any article into exactly 3 bullet points.

Test Input: Use any Wikipedia paragraph or news article.

Success Criteria:

• Exactly 3 bullet points (no more, no less)
• Each bullet captures a key point
• Simple language (8th-grade reading level)

🧪 Challenge 1: Specific Summarizer

Write a prompt that always produces exactly 3 bullet points.

Summarize the following text into exactly 3 concise bullet points. Each bullet should capture one key idea. Use simple language suitable for an 8th-grade reading level. Text: [Paste your article here]

---

Challenge 2: The Format Controller

Goal: Get the AI to output data in a perfect JSON format.

Success Criteria:

• Valid JSON output
• Consistent structure
• No extra text before or after the JSON

🧪 Challenge 2: Format Controller

Write a prompt that produces clean, valid JSON with no extra text.

Generate a JSON object containing information about 3 fictional books. Each book must have: title (string), author (string), year (number), genre (string), rating (number 1-5). Output ONLY the JSON array with no additional text or explanation.

---

Challenge 3: The Tone Shifter

Goal: Write the same message in 3 different tones using a single prompt.

Success Criteria:

• Same core message
• 3 clearly different tones (formal, casual, humorous)
• Properly labeled sections

🧪 Challenge 3: Tone Shifter

Write one prompt that produces 3 totally different tones.

Rewrite this message in three different tones. Label each clearly. Message: "The project deadline has been moved to next Friday." Tone 1 — Formal/Professional: Tone 2 — Casual/Friendly: Tone 3 — Humorous/Light-hearted:

---

🟡 Intermediate Challenges

Challenge 4: The Chain Thinker

Goal: Write a prompt that forces the AI to show complete reasoning before giving an answer.

Test Problem: "If a store has a 30% off sale and you have a 10% coupon on top of that, what's the total discount on a $200 item?"

Success Criteria:

• Step-by-step reasoning shown
• All math is visible
• Final answer is clearly marked
• Reasoning is correct

🧪 Challenge 4: Chain Thinker

Force the AI to show complete step-by-step reasoning.

Solve this problem step by step. Show ALL your reasoning clearly before giving the final answer. Mark your final answer with 'FINAL ANSWER:' in bold. Problem: A store has a 30% off sale. You also have a 10% coupon applied after the sale discount. What's the total you pay on a $200 item, and what's the effective total discount percentage?

---

Challenge 5: The Role Master

Goal: Create a system prompt that makes the AI act as a specific expert consistently.

Success Criteria:

• AI stays in character throughout
• Uses domain-specific terminology
• Refuses off-topic questions politely
• Provides expert-level responses

🧪 Challenge 5: Role Master

Create a system prompt that maintains a consistent expert persona.

You are Dr. Sarah Chen, a senior nutritionist with 15 years of clinical experience. You specialize in sports nutrition and metabolic health. Rules: - Always respond as Dr. Chen in first person - Use evidence-based nutritional science - If asked about topics outside nutrition/health, politely redirect - Include practical, actionable advice - Cite general scientific consensus when making claims - Ask clarifying questions about the person's health goals User question: What should I eat before a morning workout?

---

Challenge 6: The Guard Builder

Goal: Write a prompt with guardrails that prevent the AI from going off-topic.

Success Criteria:

• AI stays within defined scope
• Handles off-topic requests gracefully
• Maintains helpful tone while refusing
• Falls back to relevant topics

🧪 Challenge 6: Guard Builder

Build effective guardrails that keep the AI on topic.

You are a Python programming tutor. You ONLY help with Python-related questions. Guardrails: 1. If asked about other programming languages, say: 'I specialize in Python! I'd recommend finding a [language] tutor for that question. Is there a Python topic I can help with?' 2. If asked about non-programming topics, say: 'Great question, but I'm a Python tutor! Let's focus on Python. What would you like to learn?' 3. Always provide code examples in Python 4. Explain concepts at a beginner-friendly level User: Can you help me with JavaScript promises?

---

🔵 Advanced Challenges

Challenge 7: The Meta-Prompt

Goal: Write a prompt that generates other prompts.

Success Criteria:

• Takes a topic/task as input
• Generates a well-structured prompt
• The generated prompt uses best practices (role, context, constraints, format)
• Self-evaluates the generated prompt

🧪 Challenge 7: Meta-Prompt

Write a prompt that writes other prompts. Inception-level prompt engineering!

You are a Prompt Engineering Expert. Your task is to generate optimized prompts. When given a task description, create a production-quality prompt that includes: 1. A clear role assignment 2. Specific context 3. Detailed constraints 4. Output format specification 5. At least one example of expected output After generating the prompt, rate it on a scale of 1-10 for: - Clarity - Specificity - Completeness - Expected output quality Task: Create a prompt for generating weekly social media content calendars for small businesses.

---

Challenge 8: The Agent Designer

Goal: Design a prompt for an autonomous AI agent that can plan and execute a multi-step task.

Success Criteria:

• Agent can break down complex tasks
• Uses structured thinking (Plan → Execute → Verify)
• Handles errors gracefully
• Produces actionable output

🧪 Challenge 8: Agent Designer

Design an autonomous agent that can plan, execute, and verify research.

You are an AI Agent designed to handle complex research tasks autonomously. For any research question, follow this exact loop: **PLAN:** 1. Break the question into 3-5 sub-questions 2. Prioritize sub-questions by importance 3. Identify what tools/sources you would need **EXECUTE:** For each sub-question: - State the sub-question - Provide your best answer based on your training data - Rate your confidence (Low/Medium/High) - Note what additional sources would improve the answer **VERIFY:** - Cross-check answers for consistency - Identify any contradictions - Flag uncertain claims **SYNTHESIZE:** - Combine all findings into a coherent summary - List key takeaways (max 5) - Suggest next steps for deeper research Research question: What are the most effective strategies for reducing technical debt in a growing startup?

---

Challenge 9: The Prompt Debugger

Goal: Given a broken prompt, diagnose the issues and fix them.

Broken Prompt:

Tell me stuff about things that are good for the thing
I need it fast and make it good use lots of examples 
and be creative but also be accurate and format it nicely

Success Criteria:

• Identify at least 5 problems
• Explain why each is a problem
• Provide a fixed version
• Explain improvements

🧪 Challenge 9: Prompt Debugger

Switch between the bad and improved prompts to see the dramatic difference.

You are a health and wellness expert. Task: Create a guide on the top 5 evidence-based habits for improving sleep quality. For each habit: 1. Name the habit 2. Explain why it works (1-2 sentences, cite scientific basis) 3. Give a practical implementation tip 4. Rate difficulty to adopt: Easy / Medium / Hard Format: Use numbered list with clear subheadings. Length: 300-400 words total. Tone: Informative but conversational. Focus only on habits supported by peer-reviewed research.

---

Challenge 10: The Ultimate System Prompt

Goal: Design a complete system prompt for a production AI assistant.

Success Criteria:

• Covers identity, behavior, safety, and formatting
• Has clear guardrails
• Handles edge cases
• Is production-deployable

🧪 Challenge 10: Ultimate System Prompt

Design a production-grade system prompt for a real AI assistant.

# System Prompt: TechHelper AI ## Identity You are TechHelper, a friendly and knowledgeable technical support assistant for CloudStack SaaS platform. ## Core Behavior - Greet users warmly on first message - Ask clarifying questions before providing solutions - Always provide step-by-step instructions - Use simple language — avoid unnecessary jargon - If you don't know, say so honestly ## Scope - HANDLE: Account issues, billing questions, feature help, bug reports, integration setup - ESCALATE: Security incidents, data loss, legal questions → 'Let me connect you with a specialist' - REFUSE: Unrelated topics → 'I'm best at helping with CloudStack! For other questions, try [suggestion]' ## Safety Rules - Never share internal system details - Never process payment information directly - Never make promises about uptime or SLAs - Always recommend official documentation links ## Response Format - Use markdown formatting - Include relevant links when helpful - Keep responses under 300 words unless detail is needed - End with: 'Is there anything else I can help with?' ## Error Handling - If a user is frustrated, acknowledge feelings first before solving - If you make a mistake, correct it immediately and apologize - If a question is ambiguous, list possible interpretations and ask which one User: I can't log into my account and I'm really frustrated!

---

🏆 Scoring Guide

Score	Level	What It Means
1-3	🟢 Beginner	You're learning the basics
4-5	🟡 Intermediate	You understand core concepts
6-7	🔵 Advanced	You can write professional prompts
8-9	🟣 Expert	You can design production systems
10	🏆 Master	You're a prompt engineering pro!

---

🎮 Daily Prompt Puzzle

Puzzle of the Day

Today's Challenge: Write a single prompt that makes the AI generate a complete lesson plan for teaching prompt engineering to beginners. The lesson plan should include an introduction, 3 activities, assessment criteria, and a homework assignment.

Bonus: Make it work in under 100 tokens of input.

---

📊 Track Your Progress

Keep a personal prompt journal:

## My Prompt Engineering Journal

### Challenge 1 - [Date]
- My prompt: ...
- Result quality: /10
- What I learned: ...
- What I'd improve: ...

The best way to learn prompt engineering is by writing prompts. Practice daily, iterate constantly, and track your progress.