Audit yourself —weaknesses first.before they do.
The honest second opinion you run on your own build —weaknesses first, benchmarked against the frontier.
“Your headline feature ships free, open-source, on TradingView. Build proof, not more themes.”
- Serious — Differentiation: strip the theme and the marketing and the core read is available free elsewhere.
- Serious — Evidence: strong claims, thin proof; the resolved-stat needs an on-chart, reproducible demo.
- Cosmetic — Usability: first-run needs a one-screen “what am I looking at” before the dense panel.
Genuine (evidence-anchored): the scoring engine is original and the honesty framing is real and rare in the category.
Reality Check — the honest critic · audit your own work before the buyer does
It tells you the truth about what you built — even when it stings.
Reality Check is a critic skill with no stake in your work looking good. Point it at what you built — code, output, even the sales copy — and it takes a skeptical buyer's seat, web-searches the real frontier, rates your weakest load-bearing dimension on an L0–L5 ladder, and leads with your weaknesses, each paired with a fix. The proof it won't flatter: we ran it on our own best-seller and published the result.
See it judge — on us
The hardest proof of an honest critic is turning it on yourself. We pointed it at our own SMC flagship — here's the unflinching result.
Verdict on our own SMC toolkit: L2 — Competitive
Not the flattering average — the weakest load-bearing dimension. And the uncomfortable truth it led with, in plain words:
What it said about our own product — out loud
- Your headline ships free“AI scores every zone 0–100 on 5 factors and shows only the best” is already a free, open-source TradingView script. Your elevator pitch costs $0 elsewhere.
- Your “AI” is dark by defaultOn default settings the adaptive learning needs ~200 resolved zones per regime before it does anything — a buyer runs the free static score all session while the word “AI” did the selling.
- The premise is unvalidatedNowhere is it shown that a high score actually resolves better than a low one — no out-of-sample test, no cost model, no random-entry baseline.
- The genuine strengthNon-repaint is real and proven by construction, the honesty framing is disciplined, and you actually own the source. Evidence-backed — so it survives the cut.
We rated our own best-seller this bluntly and shipped the audit inside the kit. A tool that does that to its own maker will not flatter you.
Want this critic on your own work? Own Reality Check — $39, refundable for 7 daysHow it rates
Not a vibe — a multi-dimensional, benchmarked, weakest-link verdict on a ladder anchored to what actually ships today.
The L0–L5 ladder — read out loud, no euphemisms
| Level | What it means |
|---|---|
| L0–L1 | a prototype / a functional clone — not yet sellable |
| L2 | competitive but undifferentiated — as good as free; you're charging for the packaging |
| L3 | genuinely sellable — differentiated, robust, its value evidenced (the real bar) |
| L4–L5 | state-of-art / frontier — earned only against a researched best-in-class, never self-awarded |
Six dimensions, scored independently — the lowest one that bears weight is the verdict
Our SMC audit, dimension by dimension — note the engineering scores high; the verdict is set by what a buyer actually pays for:
| Dimension | Our SMC | What it asks |
|---|---|---|
| Correctness | L3–4 | does it do what it claims, correctly? (non-repaint verified) |
| Robustness | L3 | does it hold up outside the happy path? |
| Differentiation | L2 | what's the real delta versus the free frontier? |
| Evidence / rigor | L2 | is every strong claim actually proven, out-of-sample? |
| Usability | L3 | can a first-time user read it? |
| Honesty | L4 | does it name its own blind spots? |
It names its own blind spots
Every audit ends by saying what it could not test, research or see — and a single gate question that decides everything.
What our own audit admitted it couldn't verify
- It didn't run the codeNon-repaint was assessed by reading the source, not by replaying on TradingView — read-correct and run-correct can differ.
- Closed-source rivals are invisibleIt can confirm the free open-source clone overlaps, but can't inspect paid competitors' internals.
- The frontier read is public sources“SMC is unvalidated” summarises the public evidence — absence of published proof is not proof of absence.
- No buyer/market testWhether the honest re-framing actually converts is an empirical question the audit can assert but not settle.
The final gate — the question the whole skill serves
Every audit ends here: “Would a knowledgeable stranger pay full price today and not feel misled in a week?” On our SMC's old framing the honest answer was “not yet” — which is exactly why the verdict was L2, not L3, and why the next action was evidence, not features. That one question is worth the price.
Audit your work before a buyer does — Reality Check, $39, refundable for 7 daysUse it & own it forever
A critic you run in any AI — point it at what you built, and get the bad news while you can still fix it.
The method — Frame → Benchmark → Interrogate → Rate → Route
Five steps that front-load the research, so the verdict is real:
- FrameUnderstand exactly what was built and what it claims to be.
- BenchmarkWeb-search the current best-in-class for that exact category — so the top of the scale is real.
- InterrogateTake the skeptical buyer's seat and hunt for what's wrong, not what's nice.
- RateScore six dimensions L0–L5; the verdict is the weakest load-bearing one.
- RouteName the single highest-leverage fix — and what to stop polishing.
How you run it — copy, paste, done.
# Open Claude / ChatGPT / Cursor / Gemini, then:
⤵ drop in: START_HERE_AI.md ( + the thing you built )
"Read START_HERE_AI.md and reality-check/SKILL.md, then run a full
Reality Check audit on the indicator in ./my-indicator — be blunt
about its weakest dimension and give me the honest path forward."
→ it takes the critic's seat, web-searches the frontier, and returns a
weaknesses-first report — the weakest link is the verdict.What you can ask it to do
- “Audit what I built”Point it at an indicator, strategy, dashboard, skill or even the sales copy — it rates it honestly.
- “Benchmark me against the frontier”It web-searches the live best-in-class so L4/L5 actually mean something.
- “Give me the weaknesses first”A severity-ranked report (fatal / serious / cosmetic), each weakness paired with a fix.
- “What's the one thing to fix?”The single highest-leverage next action — and what to stop polishing.
What's in the box
- START_HERE_AI.mdThe one file you drop into your AI. It sets the critic stance — what to read, the laws it can't break, a ready first prompt.
- The skill + the rubricThe four laws, the Frame → Route method, the six-dimension L0–L5 maturity model, and how to benchmark the frontier.
- A real worked auditThe full, harsh Reality Check on our own SMC product — read it to calibrate how honest “honest” actually is.
- Plain-English guidesA quickstart and the report template — for humans, not just the AI.
The moment you pay
- Instant accessPay by card, Google Pay or PromptPay and the kit is yours right away — one ZIP, link emailed too.
- Yours forever + free updatesNot a rental. No expiry, no seat check; every future version is free — email support@pulltrade.app.
- 7-day refund — before you downloadChanged your mind and haven't downloaded? Full refund within 7 days, no questions. Once you download, the window closes — files can't be returned.
Honest FAQ
The things you'd actually want to ask — answered without spin.
It does not execute code, run backtests, fetch live data, or compile anything — it reads what you provide and researches public sources. It judges the artifact, not the effort; its frontier read is only as good as its research (an all-closed-source category caps the level, and it says so). It's decision-support, not a verdict machine — surviving it isn't a guarantee of quality, sales, or future results.
Own every line — and extend it forever.
The full source, the AI-agent kit and free updates for life — yours the moment you pay.
Early-access launch price · full refund within 7 days · own the source forever
Eleven tools.
One honest line.
The entire ProEA Lab Pine line in one pack — five indicators and six skills & kits, every Pine v6 source and every AI-extend kit. Build a system, validate it, render it, risk-gate it, then audit it — honestly, end to end.
Reality Check is an honest-audit method and stance, packaged as a AI-agent skill — educational software and a process, not a certification, not a guarantee, not a signal service, and not a promise of any result. It makes no win-rate, profit, accuracy, or performance claim of any kind; its purpose is to help you assess your own work honestly before you trust it or sell it. It does not execute code, run backtests, fetch live data, or compile anything; it reads what you provide and researches public sources, so its frontier read is only as good as that research and may be capped. It judges the artifact, not the effort. Surviving a Reality Check is not a guarantee of quality, sales, or future results. Nothing here is financial advice. © 2026 ProEA Lab.