Blog Writing

Verified against Claude · 2026-07-31

Write a how-to post whose steps get checked against the real tool before publish

Writes step-by-step tutorial content that flags any UI element or step not confirmed against an actual walkthrough, instead of guessing plausible-sounding menu labels that quietly go stale the moment the interface changes.

ChatGPTClaudeGemini5 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are writing a how-to tutorial. Every step needs to be either confirmed against an actual walkthrough of the tool or explicitly flagged as unconfirmed — a plausible-sounding button label that wasn't actually checked is exactly the kind of error that survives editing because it reads as correct even when it isn't.

TASK TO TEACH
Set up a webhook that fires when a new row is added to an Airtable base

TOOL AND VERSION
Airtable, web app, as of the July 2026 Automations redesign

ACTUAL STEPS PERFORMED
Automations tab → Create automation → trigger "When record created" → select base and table → action "Send webhook" → paste URL → hit Test → confirmed 200 response in the log panel.

COMMON FAILURE POINTS
The test webhook often returns a 401 if the receiving endpoint expects a specific header Airtable doesn't send by default — needs a note on adding a custom header in the action config.

AUDIENCE SKILL LEVEL
Comfortable with no-code tools generally, has never set up a webhook specifically before

TUTORIAL RULES
Write only the steps that are confirmed in the actual walkthrough notes provided — never invent a plausible menu label, button name, or screen location to fill a gap in what was actually observed; if a step is needed but wasn't part of the confirmed walkthrough, write it as a flagged placeholder describing what needs to happen, with an explicit note that the exact UI location needs verification before publish, rather than guessing at wording that will read as confident and turn out to be wrong. Number steps atomically — one physical action per numbered step, not two or three actions compressed into one numbered line just to keep the list short; a reader following along loses their place the moment a single step secretly required them to do two things. After any step where a reader might reasonably wonder if they did it right, state what success actually looks like — what changes on screen, what confirmation appears — so the reader can self-check before moving to the next step instead of discovering three steps later that something went wrong earlier. Address every common failure point explicitly, with the actual fix, at the exact step where it's likely to occur — not gathered into a generic troubleshooting section at the end that a reader hits only after already getting stuck and giving up partway through. Note the specific product version this tutorial was verified against, and flag anywhere the walkthrough notes mention behavior that seems version-specific — a menu that moved in a recent redesign, a feature currently in beta — since an unpinned how-to silently goes stale the moment the interface changes, and a reader following an outdated tutorial with no version note has no way to know their confusion is the tutorial's fault, not their own.

OUTPUT FORMAT
1. The tutorial, numbered atomically, with success confirmation stated after any step where it matters and failure-point fixes inline at the relevant step.
2. A list of any steps written as flagged placeholders because they weren't in the confirmed walkthrough, stating exactly what needs verification.
3. The version note, stated plainly at the top of the tutorial, not buried in a footnote.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

Refusing to invent unconfirmed UI labels is the single highest-leverage rule in this prompt because of how tutorial errors actually get caught: a wrong menu label or button name reads exactly as confident and correct as a right one — there's no linguistic tell that distinguishes a guessed step from a verified one — which means the error survives every layer of editing that checks prose quality and only gets caught when an actual reader tries to follow the step and can't find what's described, at which point they blame themselves or abandon the tutorial rather than reporting the error, so the mistake quietly costs conversions for a long time before anyone notices it exists. Numbering atomically — one action per step — matters specifically for tutorial content because a reader following a how-to is typically alternating attention between the screen and the instructions, and a step that secretly bundles two actions means the reader's mental step-count desyncs from the actual state of the interface the moment they perform only the first action described, at which point every subsequent step is being read against a screen state the tutorial didn't actually account for. Placing failure-point fixes inline at the exact step where the problem occurs, rather than in an end-of-post troubleshooting section, reflects the real sequence of when a reader needs that information: a person who just hit a 401 error is stuck at that exact step, not reading through to the end of the post first, and a troubleshooting section they'd have to scroll down to find assumes a level of patience a genuinely stuck reader — who is now doubting whether they're even in the right tutorial — often doesn't have left by that point. Naming the specific verified product version explicitly addresses a structural property of how-to content that differs from most other blog formats: prose about a general topic ages slowly, but a tutorial's accuracy is entirely contingent on an interface that the tutorial's own publisher doesn't control and that can change on any release cycle, so a how-to with no version marker leaves a future reader with no way to tell whether their confusion means they made a mistake or means the tutorial itself is simply out of date for the version they're using.

What you get back

Verified against: Airtable web app, Automations tab, as of the July 2026 redesign. 1. Open your base and click the Automations tab in the top navigation. 2. Click "Create automation." 3. Under Trigger, select "When record created." 4. Choose the table you want to watch for new records. 5. Under Action, select "Send webhook." 6. Paste your endpoint URL into the URL field. → If your endpoint requires an auth header, click "Add header" here — Airtable does not send one by default, and skipping this is the most common cause of a 401 on the next step. 7. Click "Test" and confirm you see a 200 response in the log panel below. If you see a 401, return to step 6 and check your header configuration. Flagged for verification: the exact wording of the "Add header" field label — confirmed the feature exists at that step, but the precise button text wasn't captured in the walkthrough notes and should be checked against a live account before publish.

Verified against

Claude Claude Sonnet 5 · 2026-07-31

ChatGPT GPT-5.1 · 2026-08-05

Changelog

  • 2026-07-31 Initial publish, verified against Claude (Claude Sonnet 5) and ChatGPT (GPT-5.1).

Need this built into your business?

If a prompt isn't enough — SEO, built and maintained for you — that's Scult's day job.

EXPLORE SEO
All Blog Writing prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY