GraphedMinds
The Startup Ideas Podcast

The Startup Ideas Podcast

The best businesses are built at the intersection of emerging technology, community, and real human needs.

Back to Frameworks

The 10 Rules of Claude Prompting Excellence

Reusability

A systematic approach to prompting Claude AI that combines tone, structure, constraints, and advanced techniques to achieve 10× better results than standard prompting methods.

How It Works

Each rule addresses a specific aspect of human-AI communication: establishing proper tone, providing clear structure, adding creative constraints, managing complexity, controlling output format, providing context, managing verbosity, offering guidance templates, using trigger phrases, and breaking down complex tasks.

Components

1

Rule 1: Use friendly, clear, firm tone (not overly polite or demanding)

2

Rule 2: Be explicit with action verbs, quantities, and target audiences

3

Rule 3: Define boundaries and constraints for more creative results

4

Rule 4: Draft outline first, then refine, then execute

5

Rule 5: Request specific output formats (markdown, tables, bullet points)

6

Rule 6: Always explain the 'why' behind your request

7

Rule 7: Control verbosity with explicit length commands

8

Rule 8: Provide templates and structural examples

9

Rule 9: Use power phrases like 'think step by step' and 'adopt persona of expert'

10

Rule 10: Break complex tasks into logical subtasks, then synthesize

When to Use

When you need high-quality, specific outputs from Claude for any task - writing, analysis, coding, creative work, business planning, or research. Especially valuable for complex, multi-step projects.

When Not to Use

For very simple, one-off questions where the overhead of structured prompting exceeds the value gained. Also not needed for basic factual lookups.

Anti-Patterns to Avoid

Being overly polite or chatty in promptsGiving vague, open-ended requests without constraintsTrying to get perfect final output in one promptNot specifying desired output formatSkipping context about why you need somethingNot controlling output length explicitlyAsking for complex deliverables without breaking them down

Example

Instead of 'Tell me about stoicism,' use: 'Act as a university professor of philosophy. I'm preparing a 1-hour intro lecture for students with no prior knowledge. First create a lecture outline with three main sections. The outline should have clear introduction, body, and conclusion. Please format this as a nested bulleted list. For each major point, include a key stoic figure and one of their core ideas. Your tone should be accessible and engaging.'