Launched this week
Vibe-check skill
Stop your AI from building the wrong app
26 followers
Stop your AI from building the wrong app
26 followers
Every coding skill starts after you've decided what to build. vibe-check covers the part before, the part that sinks most projects. It digs out the real problem under your vague idea, pressure-tests it against what people actually gripe about on Reddit, and hands your AI agent a blueprint to follow. It can even tell you no, a real no-go verdict. Other skills help you build it right. This one makes sure you're building the right thing.
RiteKit Company Logo API
@amer_arab Everyone's shipping coding skills for after you've decided what to build — covering the part before, digging the real problem out of a vague idea and pressure-testing it, is where most projects actually sink. "Stop your AI from building the wrong app" names a pain a lot of people feel but couldn't quite put into words.
The dig-out-the-real-problem step is best shown in motion, and you launched without a demo — so I made you one, free and whitelabel, no strings:
Yours to keep — download it from https://foxplug.com/v/ss-vibe-check-launch-from-a-vag-425ab013 and put it on your own channel or launch page. Launches with a video do better, and yours is still editable.
Made at https://foxplug.com/?utm_source=producthunt&utm_medium=comment — make more there, or record your own product tour in ~2 minutes. Anyone else launching soon: paste your site, video in about 30 seconds. Nice launch.
@saulfleischman Thank you so much for this. I appreciate it.
RiteKit Company Logo API
@amer_arab Actually, the market for your product and FoxPlug has overlap, but we're in no way competing. To be perfectly honest, for a long time we've been grinding on the auto-generated launch videos - where we're working off nothing but a landing page and (from today) other materials, like media in a PH launch page. I ask you, thus, what would make the video good enough that you would include it in your media gallery - in PH launch page, elsewhere?
the no-go verdict is the part I'd stress test hardest. it's easy to feel good about killing a bad idea, but what about ideas that look weak on paper because the market signal is sparse or contrarian rather than because the idea is actually bad? how do you keep vibe-check from converging on "whatever already has visible reddit complaints" as its working definition of validated, since that would quietly bias it toward incremental ideas over genuinely new ones nobody's complained about yet.
@galdayan I say that I'll get you 80% of the way there, but the 20% is still gut feeling. I'm a big believer in gut feeling. The benefit of using this skill is that it will at least expose all of the evidence for you, for your idea and against it. But then you have to decide yourself.
Oh, and by the way, I don't rely only on Reddit. There are multiple other sources like G2 reviews and app store reviews. The vibe check skill kind of look for the general sentiment all over the internet, not just Reddit.
@amer_arab good to know it's not reddit-only, that changes the bias concern a bit. still curious how the scoring handles conflicting signal though - like one sharp, detailed 1-star review that nails a real gap versus fifty generic 5-star ratings with no substance. does the score weight specificity/detail over raw volume, or does it end up averaging toward the noisy majority?
@galdayan First, it scores how big each problem is and tags how solid the evidence is, and those stay two separate numbers. A problem can rank at the very top of the list and still get marked "hunch" instead of "seen it," because fewer than three independent sources backed it up. You see the size and the confidence side by side. One never hides the other.
Second, the evidence gates the call, not the score on its own. If the highest-scoring problem is only a hunch, the skill won't hand you a green light. It makes you go pull one more thread on that specific problem, or demote it on the record and say out loud why. That's the guard against getting too optimistic, which is the direction that scares me more.
When it's genuinely a coin-flip, the output isn't a rating at all. It's a conditional plus the cheapest next move: narrow to a sharper audience and run it again, aim at the adjacent problem the evidence actually points at, or stop and walk with a findings summary that lists every open question next to the small test that would answer it. Ten DMs. A landing page that costs you an afternoon instead of a month.
So a borderline case resolves into something like "here's what's solid, here's what's still a bet, and here's how to settle the bet before you write a line of code."
And the honest part, the one I put right in the README: this gets you maybe 80% of the way. The skill filters the noise and checks every quote against a real source, so you're not scoring a hallucination. The last stretch on a true coin-flip is a human call, and I think it should stay one. Ideally made by the person who actually lives inside the problem. I'd rather hand you a clear-eyed picture and trust your gut on the margin than fake a verdict the evidence can't hold up.
@amer_arab the honesty about the 80% and human-call-on-the-margin part is what makes this credible, most people selling a scoring tool wouldn't admit the ceiling. quick question on the "three independent sources" bar though - how do you catch sources that look independent but aren't, like five reddit threads that are all actually just quoting the same original viral post or complaint. that would count as three-plus and get promoted to "seen it" when really it's one person's opinion wearing five hats
I like that you're separating "build the product well" from "build the right product."
As AI makes software creation cheaper, I think the biggest source of failure shifts upstream. Building becomes abundant, while deciding what deserves to be built becomes the scarce capability.
That feels like the more important problem to solve.
@aryan787544 Totally! if you've been on reddit recently every subreddit about building a new product is people complaining that they've built something and that they spent weeks building it but there are no users for it. A lot of people are building the wrong thing without knowing. Vibe check skill will significantly increase the odds of you building the right thing.
@amer_arab I think that's the subtle shift that's happening.
A lot of founders assume the problem is "validation," but it's often interpretation. They collect plenty of signals—they just don't know which ones have actually earned the right to influence the product.
If AI can help founders distinguish noise from genuine evidence before they start building, that's a much harder problem to solve than simply finding more feedback.
@aryan787544 And in addition to validation, there is differentiation as well. Because when everyone can build, it means you really have to be unique in order to succeed and really solve the problem, the right problem really well. And that's exactly what VibeCheck helps you to do.
@amer_arab I think that's where the next challenge begins. If AI helps founders avoid building the wrong product, the bottleneck shifts from deciding what to build to deciding which evidence actually deserves to shape the product.
I'm curious: what outcome would convince you that Vibe-check is improving founders' judgment rather than simply giving them more confidence? That distinction feels like it could become your strongest differentiator.