A FIELD GUIDE FOR WORKING WITH CLAUDE
Move from a blank chat to a clear next step. A practical collection of 120 prompt patterns for the writing, coding, and decisions you make every day.
One payment Lifetime updates Yours to keep
“Make this sound more human.”
Edit the draft below for [audience] who needs to [take a specific next step]. Keep the meaning and factual claims within the supplied source material. Voice reference: [paste a short sample of your own writing that you have permission to use] Prefer concrete nouns, direct verbs, and sentences I would say aloud.
Read the complete promptAn editable starting point. Results depend on your task.
THOUGHTFUL TOOLS. EVERYDAY WORK.
A reference you can actually use.
Choose a task, adapt a pattern, and check the result against your own work.
Prompt patterns organized around what you need to do, with worked examples and notes on when to leave a pattern out.
Start with a useful question.
Then give it the context it deserves.
Playbooks and combo templates help you connect individual prompts into a sequence you can revisit.
Try three free workflowsA SIMPLE PRACTICE
Find a pattern for the task in front of you.
Replace the placeholders with your goals and constraints.
Review the output. Keep what helps. Question what does not.
Try two before you buy · No email required
Every one of the 120 patterns looks like this: what it does, the exact prompt, the recorded before and after, and when to skip it. Copy either prompt, swap in your own task, and see the difference in Claude right now.
Tells Claude to challenge the premise of your question before answering it. Most prompt prefixes change how Claude answers. /skeptic changes which question Claude attempts to answer in the first place. That distinction is what makes it the only code I would defend as a real logic intervention rather than a structural tweak.
/skeptic Should I add a referral program to my SaaS where buyers earn 30% commission for every new buyer they refer?
Should I add a referral program to my SaaS where buyers earn 30% commission for every new buyer they refer?
Generic Claude: A 400-word writeup of how to structure a referral program — commission tiers, attribution windows, payout cycles, fraud prevention. Useful if you should build a referral program. Disastrous if you shouldn't.
"Before I answer the structural question, I want to challenge the premise. How many existing customers do you have? If the answer is under 50, a referral program is almost certainly premature. The signal-to-noise on referrals is too low at small scale, and you'll spend more time building the program than it pays back in qualified referrals. The right move at <50 customers is to personally ask each existing customer for one specific intro, not to build a system. If you're over 50 customers, here's how I'd structure it..."
The first version answers the question. The second version catches the wrong question before answering, then answers the right one if it survives the challenge. In the test, this exact prompt was run by a founder with 4 customers. The referral program would have been a weekend wasted.
Factual questions ("what year was X founded") — adds friction with no upside
Strips the structural patterns that make AI writing recognizable as AI writing. Specifically: em-dashes used for rhetorical pauses, transitional phrases that mark essay structure, balanced sentence pairs (one short sentence followed by one long sentence repeated as a pattern), and meta-commentary like "I hope this helps" or "let me know if you need clarification." The output reads like a tired human typed it at 11pm — slightly less polished, much more believable.
/ghost Write a cold email to a SaaS founder about my product.
Write a cold email to a SaaS founder about my product.
Generic Claude: "Dear [Founder], I hope this email finds you well. I wanted to reach out to introduce you to our innovative platform — a comprehensive solution that streamlines your sales workflow. Furthermore, our product has been carefully designed to integrate seamlessly with your existing tools. I would love to set up a quick 15-minute call to discuss how we can help you achieve your goals. Looking forward to your response! Best regards, [Name]" — every sentence is an AI tell. The recipient deletes it before finishing the first paragraph.
"Hey [Founder], saw your post about response rate plateaus last week. Built a thing that might help — it loads ten distinct sales philosophies (Collison, Benioff, etc.) and writes outreach in their voices instead of generic AI tone. Free version at [link], $599 for the full pack. Reply with 'send it' if you want details. — [Name]" — no em-dashes, no transitional phrases, no "I hope this finds you well," no "looking forward to your response." Reads like a real person wrote it.
The structural difference is enormous and the audience that matters can spot it instantly. Modern technical buyers have spent two years training themselves to recognize and delete AI-generated emails. /ghost is the single most important defense against that filter.
Technical documentation that needs to sound authoritative — /ghost makes it sound casual and slightly broken on purpose
Bought it mostly out of curiosity. The placebo verdicts section alone was worth it. I was using three prompts that apparently do nothing.
Used two of the combo-stacking prompts on a client deliverable this week and it saved real time. This is the first prompt resource I have paid for that felt tested rather than guessed.
The anti-pattern library changed how I write prompts more than the codes themselves did. Small purchase, disproportionate value.
Solid for the price. Wish there was a searchable index instead of scrolling the PDF, but the before/after examples are genuinely useful.
Good honest tone throughout, appreciated that it tells you what does not work instead of just selling hype. Delivery was instant which I did not expect for $15.
Decent reference but I expected more depth on the failure-mode warnings. Half of them are one line. Still cheaper than the hours I would have spent testing myself.
It is fine. Not life-changing, but I do not regret the $15. Would not recommend at a higher price point though, the depth does not quite match a $49 pack.
Some codes felt like stuff I could have found free on Reddit. The token-delta data is the actual differentiator, more of that and less repackaged basics would have made this better.
Felt like I should have gotten a bigger sample before judging the rest of it. Could be clearer upfront about what you get.
THE COMPLETE CLAUDE CHEAT SHEET
A compact reference for the moments when you know what you want to do, but need a better way to ask.
Start with the free guidePersonal license. Refund policy · Already purchased?
Better inputs. Thoughtful checks. Your work, a little clearer.
Try something free