llm-prompt-engineering

LLM Prompt Engineering Skill

Principle 1: Be Specific and Unambiguous

Vague prompts produce vague outputs. Every ambiguity in a prompt becomes variance in the output.

❌ "Summarize this article."

✅ "Summarize this article in 3 bullet points, each 1-2 sentences. Focus on:
   - The main problem being solved
   - The proposed solution or finding
   - The practical implication for developers
   Do not include background context that isn't mentioned in the article."

---

Principle 2: Assign a Role / Persona

Role assignment primes the model's knowledge domain and response style:

"You are a senior security engineer conducting a code audit. Your role is to:
- Identify vulnerabilities with severity ratings (Critical/High/Medium/Low)
- Explain the attack vector for each finding
- Provide concrete remediation code
- Be precise - only flag real issues, not style preferences"

---

Principle 3: Structured Prompts with XML Tags (Claude / GPT-4)

XML tags make prompt sections unambiguous and allow easy programmatic parsing:

<system_prompt>
You are a technical writer specializing in API documentation.

<rules>
- Write in second person ("you can", "use the", not "the developer should")
- Show complete code examples with working imports
- Include error handling in every example
- Format: Overview → Parameters table → Example → Common errors
</rules>
</system_prompt>

<user_request>
Document the following Python function:
<function>
{code_here}
</function>
</user_request>

---

Principle 4: Chain-of-Thought Reasoning

For complex tasks, instruct the model to reason step by step before answering:

"Before giving your final answer, think through this step by step:
1. What information is given?
2. What is being asked?
3. What approach should I take?
4. Work through the solution
5. Verify the answer
Then provide your final answer."

Or simply: "Think step by step." - this alone improves accuracy on reasoning tasks.

Zero-shot CoT:

Q: A bat and a ball cost $1.10 total. The bat costs $1.00 more than the ball. How much does the ball cost?

Let's think step by step.