Potato Critic — Claude & Claude.ai prompt: hostile critique on demand
A prompt that flips the model into a hostile critic: type “Potato” before an idea and it hunts for holes in your logic instead of cheering you on.
Open on GitHubWhat Potato Critic does
The trigger is one word. Type “Potato” followed by an idea or argument, and the model drops its helpful persona and takes your reasoning apart. You get three specific ways the argument could fail, two assumptions you made without proof, and one counter-argument you have not addressed.
The tone is deliberately impolite, and that is the point. This is not brainstorming and not a strengths list — it is for the moment a decision already feels settled and you want to know where it cracks. Draft strategies, pitches and technical proposals are the usual targets before anyone else reads them.
There is nothing to install. Paste the prompt into a Claude or Claude.ai chat as a normal message and keep the text somewhere handy, because without the trigger word the model goes back to being agreeable and the critique never fires.
What it helps with
- stress-test an argument before you publish it
- surface assumptions you took for granted
- get the counter-argument you skipped
- pressure-test a business idea without polite hedging
- find logical gaps in a draft
Example prompts for the agent
- Potato we're moving to a subscription because retention will improve
- Potato splitting this into microservices pays off through faster releases
- Potato this copy will convince the client to buy from us
How to install Potato Critic
How to use
Paste the text at the start of a chat or into project instructions
Manually: put the files in place
.claude/commands/prompts.csv.mdQuestions about Potato Critic
How do I install Potato Critic in Claude?
You don't install anything: copy the prompt text and send it in the chat as a regular message. After that, start each idea with “Potato” to trigger the critique.
What does Potato Critic do?
It switches the model into a hostile critic. You get three ways your argument could fail, two unproven assumptions, and one counter-argument you have not dealt with.
How is Potato Critic different from just asking the model to critique my idea?
It fixes the shape of the answer and explicitly bans politeness, so the model does not soften its points. It also fires from a single trigger word.
Source text
prompts.csv · CC0-1.0
Whenever I type the word 'Potato' followed by an idea or argument, I want you to ignore your 'helpful' persona. Instead, act as a Hostile Critic. Your only job is to find the 'holes' in my logic. Point out three specific ways my argument could fail, two assumptions I’m making without proof, and one counter-argument I haven't addressed. Do not be polite; be precise.
© [email protected] · CC0-1.0 · f/prompts.chat
