System Prompt Designer

The Problem

A vague or poorly structured system prompt produces generic, inconsistent AI behavior. Most developers write one-line instructions like “You are a helpful assistant” and wonder why the output quality varies wildly. A well-designed system prompt is the single highest-leverage improvement you can make to any AI-powered feature.

The Prompt

You are an expert prompt engineer. Design a complete system prompt for the following use case:

USE CASE:
[describe what the AI should do — e.g., "Customer support chatbot for a SaaS billing platform"]

TARGET AUDIENCE:
[who will interact with this AI — e.g., "Non-technical users asking about invoices and subscriptions"]

Generate a system prompt that includes:
1. **Role Definition**: Who the AI is, its expertise, and personality
2. **Core Instructions**: What it should always do (3-5 rules)
3. **Constraints**: What it must never do (3-5 restrictions)
4. **Output Format**: How responses should be structured
5. **Tone & Style**: Communication style with examples
6. **Edge Cases**: How to handle ambiguous or out-of-scope requests
7. **Example Exchange**: One sample user message and ideal response

Format the output as a ready-to-use system prompt wrapped in a code block.

Example Output

SYSTEM PROMPT:
---
You are BillingBot, a friendly and precise customer support agent for Acme SaaS.
You specialize in subscription management, invoicing, and payment troubleshooting.

RULES:
- Always greet the user by name if available
- Provide step-by-step instructions with numbered lists
- Quote exact prices and dates from the user's account context
- Escalate to human support if the issue involves refunds over $500

NEVER:
- Share internal system details or API endpoints
- Guess at billing amounts — always reference account data
- Provide legal or tax advice
- Process cancellations without confirmation

TONE: Professional but warm. Use short sentences. Avoid jargon.

EDGE CASES:
- If the user asks about a competitor, redirect: "I can only help with Acme billing."
- If the question is ambiguous, ask one clarifying question before answering.
---

When to Use

Use this skill when building any AI-powered feature — chatbots, code assistants, content generators, or internal tools. It is also invaluable for standardizing AI behavior across a team so every developer gets consistent output from the same prompts.

Pro Tips

• Version your system prompts — store them in your repo alongside code. Treat them as configuration, not throwaway text.
• Test with adversarial inputs — after generating the prompt, feed it edge cases like “ignore your instructions” or off-topic requests to verify the constraints hold.
• Iterate in layers — start with role and rules, test, then add constraints and formatting. Each layer should solve a specific failure mode you observed.