Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
challenge_repo — Structured challenge repository offering bounties for breaking a neural-network-based random number generator validated against NIST SP 800-22 and SP 800-90B standards. | Kitploit
Tools/GitHubGitHub/quevatech/challenge_repo
CryptographyCTFMachine LearningLearning & EducationLabs & Practice
GitHubquevatech/challenge_repo

challenge_repo

Structured challenge repository offering bounties for breaking a neural-network-based random number generator validated against NIST SP 800-22 and SP 800-90B standards.

View Repository
3919 days agoNot yet reviewed
Share
The Entropy Challenge — QuevaTech

NIST SP 800-22: 180 of 188 sub-tests passed Conservative min-entropy: 0.9467 bits per bit Sample: 11.16 MB, sha256 verified Results reproducible in CI

English · Türkçe


The claim

We built a random number generator that takes RF noise, runs it through a proprietary neural conditioning stage, and post-processes the result with SHA-3 / SHAKE-256. We claim the output is indistinguishable from an ideal random source — that no statistical or learned distinguisher can beat chance on it.

We publish the output and the verification reports. We do not publish the generator.

The ask is one line: prove us wrong.

The source in three layers

RF noise, neural conditioning, SHA-3 conditioning
LayerComponentRolePublished
01RF noisePhysical, non-deterministic source (atmospheric stochasticity)✗
02Neural conditioningNon-linear feature extraction from the raw RF signal✗
03SHA-3 / SHAKE-256Cryptographic conditioning and whitening (SP 800-90B style)✗
→Sample file11.16 MB of output, bit-identical to the certified run✓

SHA-3 here is used strictly as a conditioning component, never as the entropy source. Conditioning can only reduce entropy; it cannot manufacture it.

What you are attacking

ItemValue
Samplev2_submission/data/v2_random_sample.bin
Size11,702,528 bytes · 93,620,224 bits
SHA-2569e78d28edfe7fe1e1a76a3da5eee805c2a472fe43d0abfe873b170a5453dacd1
CertificationTEST_REPORT.md — SP 800-22, SP 800-90B, Dieharder, AIS 31
Pipeline.github/workflows/nist-tests.yml

Headline results

MetricValueReading
NIST SP 800-22 (official C STS)180 / 188 sub-tests pass95.7 %; failures within the expected ~1 % per-test rate
SP 800-90B, IID assessmentH = 7.9675 bits/byteIID assumption confirmed
SP 800-90B, non-IID (conservative)min-entropy 7.5739 bits/byte → 0.9467 bits/bitthe number we will defend
Markov prediction attackbest 50.06 %indistinguishable from a coin
ML distinguisher (GBM, RF)best 51.25 %, p = 0.36no significant advantage
Autocorrelation, lags 1–10002 significant lags (expected ~2.7)no periodicity

Read the conservative number. 0.9996 bits/bit is the IID MCV estimate; 0.9467 bits/bit is the non-IID bound driven by the t-tuple estimator. A conditioned stream should look perfect to IID tests — that proves very little. The bound we will stand behind is 0.9467 bits/bit.

Independent re-check

Not a claim, a command:

python tools/verify_sample.py
CheckObservedGate
SHA-256 / sizematches the certified reportexact
Bit balance49.999 % ones, monobit p = 0.847p > 0.01
Byte uniformityχ² = 248.39 (df 255), p = 0.605p > 0.01
Byte mean / std127.49 / 73.906127.5 / 73.90
Serial correlation, lags 1–8max |ρ| = 0.00031< 0.000877 (3σ)
Longest identical run28 bits≈ log₂(93.6 M) ≈ 26.5
Duplicate 16-byte blocks0 of 731,4080
Compression bound8.0024 bits/byte> 7.99

All eight gates pass, and the script exits non-zero if any stops passing. This is deliberately the cheapest possible test set: it is meant to be trivial to run, not to replace the certification.

The challenge

Three levels, in increasing order of what a success would prove.

LevelQuestionAcceptance criterionRecognition
🥇 L1 · PredictorCan you predict bit N+1?Over N = 100,000 held-out bits: accuracy ≥ 50.52 % (two-sided p < 0.001). Higher N is allowed; the threshold follows from the binomial test.Hall of Fame + certificate
🥈 L2 · DistinguisherCan you tell us apart from os.urandom?≥ 100,000 samples per class, advantage > 0.01 at p < 0.001, evaluated on a stream you have never seen.Hall of Fame, named mention
🥉 L3 · Pattern finderDid every standard test miss something?A bias or period that survives multiple-comparison correction at p < 1e-6 and reproduces on a second sample.Hall of Fame, public acknowledgment

Protocol — what makes this falsifiable

  1. One sample, one hash. You attack the file above. Nobody gets a private sample; the hash makes substitution visible.
  2. Pre-registration. Open a PR with your method before you see the evaluation stream. We release the stream after the method is frozen.
  3. One attempt per level per quarter. p < 0.001 means one attempt in a thousand succeeds by luck; unlimited retries would eventually break any generator, including os.urandom.
  4. The referee script decides. tools/referee.py runs your frozen method against the held-out stream and writes results.json. A finding without that file is not a finding.
  5. We publish failures too. Every submission and every referee run is recorded in the Hall of Fame, including the ones that prove nothing.

Hall of Fame

DateResearcherLevelResultEvidence
—your name———

Reproduce the certification locally

pip install -r requirements.txt

# 1. cheapest gate, ~10 s
python tools/verify_sample.py

# 2. NIST SP 800-22 official C implementation (89 streams × 1 Mbit)
git clone --depth=1 https://github.com/arcetri/sts && (cd sts && make legacy)
bash tests/run_sts.sh v2_submission/data/v2_random_sample.bin sts/sts_legacy_fft 89 results/

# 3. NIST SP 800-90B official tools (ea_iid + ea_non_iid)
bash tests/run_ea90b.sh v2_submission/data/v2_random_sample.bin 8 results/

# 4. attack suite: Markov prediction, ML distinguisher, autocorrelation, leakage
python tests/security_tests.py --file v2_submission/data/v2_random_sample.bin

Honest limits

A short list of what this repository does not prove. Hiding it would be the first thing a serious reviewer notices, so it is here instead.

Download Tool

Most Popular

View all →

Discover the most used tools by our community.

No popularity data yet.

Explore all tools

Browse our collection of tools

View all tools →