👨‍🎓

prompt engineering guides

There is an absolute ton of guides to learn prompt engineering out there. Whole portals in fact.

Here are some of the best current prompt engineering guides:

Anthropic Claude 4 prompt engineering best practices - AnthropicAnthropic Claude 4 prompt engineering best practices - Anthropic

AnthropicAI AI Learning Resources & Guides from AnthropicAnthropicAI AI Learning Resources & Guides from Anthropic

OpenAI PlatformOpenAI Platform

OpenAI Academy Content | OpenAI AcademyOpenAI Academy Content | OpenAI Academy

Google Cloud Prompt Engineering for AI Guide | Google CloudGoogle Cloud Prompt Engineering for AI Guide | Google Cloud

The Hitchhiker's Guide to Grok | xAI DocsThe Hitchhiker's Guide to Grok | xAI Docs

Go searching, there’s a lot more.

But before you go doing a degree on it - I highly recommend checking out a project called Fabric by Daniel Miessler: https://github.com/danielmiessler/fabric.

Daniel is a talented individual who has created a collection of powerful, well-crafted prompts.

If you’re looking to learn prompt engineering, my advice is to start by studying these prompts. Take time to understand what they’re designed to do, then explore the various techniques covered in the following content. It’s often easier to work backward: begin by examining what “good” looks like, as showcased in the prompts below.

Fabric prompts

‣
analyse [16]
‣
create [33]
‣
explain [4]
‣
extract [22]
‣
find [2]
‣
get [2]
‣
improve [4]
‣
rate [5]
‣
recommend [3]
‣
summarise [11]
‣
write [6]
‣
miscellaneous [15]
fabric prompt acknowledgementfabric prompt acknowledgement

Conclusion

Its worth tothinghat you can use AI to assist with prompt engineering. Simply describe what you want in plain English, and ask the AI to craft a prompt for you. The prestige once associated with prompt engineering has diminished, but understanding how to maximise your AI models’ potential remains crucial—especially if you aim to achieve remarkable results. This is particularly relevant for tasks like AI coding or performing the precise, powerful functions we’ll explore later.

My second piece of advice: Don’t stress too much about perfecting your prompt engineering. It shouldn’t hold you back. Just experiment—ask a question, see what the model returns, and then ask it to refine the prompt. Run it again. Also, consider the data you’re providing.

You don’t need to get it right the first time. Iterate toward a great solution.