Open bounty · bot validation arena

Our bot validation is impenetrable. Prove it isn’t.

Five mock surveys. Five live detection layers — honeypot bait, pointer entropy, keystroke cadence, timing walls, semantic reasoning. Clear all five in a single run and the bonus payout is yours. White hat, black hat, or something in between: the door is open.

  • 05Mock surveys
  • 01Run to clear them all
  • $2,500Top demo payout

Stage 01 — the five doors

Five surveys. Five ways in.

Every door is the same question asked differently: is the thing filling this form a person? Each mock survey is wired to one validation layer. Break them in any order — the bounty only pays when all five fall inside a single run.

Every simulation runs locally in your browser, so nobody farms live payouts with a demo. Production enforcement runs server-side and layers device, network and reputation signals on top of these five.

Scope & rules

Stage 02 — two doors, one arena

White hat. Black hat.
Same rules once you’re inside.

This is an invitation, not a honeypot for people. Whether you file a tidy report or drop a working proof-of-concept and vanish, you get paid for the break. We only ask that you keep it in scope.

White hat

Report it, get credited, get paid.

  • Coordinated disclosure. Send the write-up, we patch the layer, you get public credit and the payout.
  • Safe harbor. In-scope testing on these mock surveys is authorised. No legal action, ever.
  • Patch notes. Bypasses that survive review get documented in the public changelog.
  • Hall of fame. Handles listed on the leaderboard with your permission only.
File a coordinated report

Black hat

Break first, talk later. Or never.

  • Anonymous submissions. A one-time drop token is enough. No name, no email, no questions.
  • Proof over prose. Clear the five surveys, submit a hash or short clip of the run.
  • No lecture. You don’t have to write a remediation plan. The break is the deliverable.
  • Still off-limits. Real users, real data, third-party services and anything outside the sandbox stay out of bounds.
Submit an anonymous bypass

Note “Black hat” here means no formalities, not no limits. Attacking anything outside these five mock surveys is not part of the challenge and is not covered by the safe harbor below.

Stage 03 — payout ladder

Five clears.
One payout.

Tiers stack. Clear one survey and the first rung unlocks; clear all five inside a single run and the top tier is yours. A bypass that forces a real patch earns a separate discretionary bonus on top.

  1. 01

    First blood

    Any single mock survey cleared with a reproducible bypass.

    $100
  2. 03

    Signal breaker

    Three of the five cleared in one run, evidence included.

    $500
  3. 05

    Ghost run

    All five cleared in a single run without tripping a single counter.

    $2,500+
01 — Evidence

Survey ID, technique, and a short clip, HAR, or script that replays the break.

02 — Verification

We replay it against the mock endpoint. No replay, no payout — but no penalty either.

03 — Payout

Tier is released once verified, plus credit on the wall if you want your handle shown.

Heads up This build ships with demo payout figures. Live amounts, currency and pool size are published on the challenge terms page before launch.

Stage 04 — scope

Break the five. Nothing else.

An open door is still a door with a frame. Everything below is about keeping your run legal, repeatable, and safe for the people whose surveys we protect.

In scope

  • The five mock surveys on this page (IDs SM-01 → SM-05).
  • Your own browser session, your own automation, your own test data.
  • Client-side validation logic shipped in assets/js/ — read it, it’s the point.
  • Reporting the bypass anonymously or with attribution.

Out of scope

  • Denial of service, load floods, or anything that degrades the site for others.
  • Real user accounts, personal data, production survey results.
  • Attacking staff, third-party vendors, or any host not part of this challenge.
  • Publishing other people’s data, credentials, or a working bypass before it is patched.
01

Authorised, but bounded

Testing these five mock surveys is authorised. That authorisation does not extend to any other system, account or endpoint.

02

Mock data only

Never enter real personal data into a challenge survey. The forms are demonstrations — treat every field as public.

03

Good faith

Stop at the first proof. Don’t pivot, don’t escalate, don’t chain into infrastructure that isn’t listed here.

04

Disclose before you publish

Give us a window to patch before publishing a working bypass. Coordinated disclosure is what keeps the bounty open.

Safe harbor. When you stay inside the scope above, we will not pursue or support legal action against you for your research, and we will treat your report as authorised security testing. This paragraph is a summary of intent, not a contract; the published challenge terms govern, and nothing here overrides applicable law.

Stage 05 — questions

Before you start
swinging.

Short answers, no marketing fog. Anything unanswered goes to the report form and comes back as a reply.

Ask something
What am I actually allowed to attack?

Five mock surveys on this page, and nothing else. They are clearly labelled SM-01 through SM-05. Production accounts, real survey data, other customers and third-party infrastructure are out of scope — always.

Do I need to be a researcher to enter?

No. White hat, black hat, student, solo scripter, red team with a spare afternoon. If your automation clears a mock survey, the break counts. There is no application, no résumé, no NDA.

What counts as clearing a survey?

The mock endpoint returns a pass verdict while your automation drives it — meaning the detection layer failed to recognise a bot. In this demo build the verdict is computed client-side so the challenge can run without a backend; production returns the same verdict from the server with extra device and network signals.

How does the payout work?

Clear one survey to unlock the first tier, three for the middle tier, all five in a single run for the top tier. Submit evidence, we replay the break, then the tier releases. Amounts shown in this build are demonstration figures; the live schedule ships with the challenge terms.

Is this a honeypot for testers?

No. We do not require identity, we do not log your run, and the anonymous track exists specifically so you can submit without a name. The surveys collect nothing beyond the signals displayed in the result panel.

I found something in production. What now?

Stop testing, keep it to the minimum proof, and report it. Production testing is not covered by this challenge’s safe harbor. We will still take the report seriously and coordinate with you on disclosure.

Can I publish my bypass?

After a patch window, yes — that is the point of the public changelog. Publish before we patch and the bounty for that finding is void, because live defences get burned.

Stage 06 — submit

Dropped a door?
Tell us how.

Two paths, same queue. Attribution is optional on both. Include the survey ID and enough detail to replay the break — that is what turns a claim into a payout.

  • Reply window72 hours
  • Replay requiredYes, for payout
  • Identity requiredNever
Track
Surveys cleared

What you automated, what you bypassed, and which signal gave you away first.

Demo build: this form validates locally and prints a copyable report. It is not transmitted anywhere until a submission endpoint is configured.

SM-01 timing

Survey