PARTLY

Partly true: the '30 ChatGPT secret codes' are ordinary prompt shorthand that works fine, and the caption admits they are not hidden commands, but the card still sells them as secret codes for '10x results'

Facebook · Oct 2026

The claimOn-screen card: '30 ChatGPT Secret Codes: The Short Codes to 10x Results', listing /ceo, /seo, /eli5, /list, /code, /human, /viral, /brief, /table, /debug, /email, /expert, /critic, /expand, /teacher, /improve, /compare, /outline, /strategy, /research, /simplify, /motivate, /summarize, /translate, /interview, /copywriter, /brainstorm, /explaincode, /coverletter and /promptengineer, each with a one-line meaning. Caption gives worked examples for /critic, /compare and /interview and says the codes are 'shorthand for instructions, not hidden settings or official commands.'

This one deserves some credit. The caption says plainly that these 'codes' are shorthand for instructions, not hidden settings or official commands, and that plain English works just as well. That is accurate. OpenAI documents no /critic, /compare or /interview commands for ChatGPT; its own slash-command reference lists different, product-specific commands such as /review and /plan, and third-party write-ups confirm the slash itself does nothing. What makes a word like '/critic' work is that the model reads it as a word. The worked examples in the caption are genuinely good prompts: ask it to flag vague promises and missing scope in a proposal, compare options against your budget and timeline, interview you one question at a time before advising. Those habits do improve answers in ChatGPT, Claude or Gemini. The trouble is the packaging. The video card, which is what most viewers see, says 'Secret Codes' and '10x Results', and lists thirty one-word labels with no context, so anyone who does not open the caption walks away believing typing '/viral' unlocks something. A one-word tag with no goal, audience or example rarely beats a clear sentence. The page is not selling anything here beyond follows and saves, which counts in its favor. See also the earlier '35 ChatGPT codes' entry, graded partly.

What holds up

Caption and contact sheet read; all 30 codes on the card transcribed.
OpenAI's slash-command reference lists /review, /side, /status, /plan and others; none of the card's 30 codes appear.
Third-party guides (Tom's Guide, Towards AI) confirm these are user prompt conventions with no built-in behavior.
Matches an earlier audit of a '35 ChatGPT codes' card graded partly for the same reason.

What doesn’t

'Secret Codes' and '10x Results' on the card contradict the honest caption.
Thirty bare labels with no context teach incantation over clear instruction.
No comment gate or paid funnel; engagement-only post.

The catch

The caption tells the truth and the card does not. The useful part is the habit of asking the model to critique, compare and interview you, not the slash.

How to actually do it

Skip the slash; write one sentence with the role, the goal and the output you want, for example 'Critique this proposal as a skeptical buyer and list the five questions I must answer.'
Save your three favorite instructions as ChatGPT custom instructions or a Claude project instruction so you do not retype them.
For big decisions, chain them: interview me, then critique, then compare two options, exactly as the caption suggests.

Use the caption, ignore the card. The prompting habits are sound and free; the 'secret code' label is the only hype.

Confidence: High
Posted by: an AI-tips page for small business (about 23,000 plays); reel sent to Buddy Tue Oct 27, 2026, ~11:03 PM

See the original claim →

We test hype for free. We build the real thing for a living.

Thirty minutes, no pitch — and you'll leave with something useful either way.

Book a call with Todd or start with the free Business Checkup →

The Verdict Weekly

Three verdicts every Friday. Free forever, unsubscribe anytime, no spam — that would be ironic.