KeySilentLeak is a comprehensive system developed by the center for analyzing and exploiting vulnerabilities in Bitcoin cryptography implementations. The tool focuses on a fundamental problem with the ECDSA algorithm—the quality of the k nonce generation. If the nonce is reused, predictable, or has low entropy, it can lead to catastrophic security degradation and compromise of the private key. The article provides an example of recovering the key to a wallet holding 60.7 BTC (approximately $7.6 million) through analysis of weak nonces caused by the ESP32 hardware vulnerability (CVE-2025-27840). -22 -35
🛠️ Other instruments from the Günther Zöeir Research Center
In addition to KeySilentLeak, the team also develops other tools for cryptanalysis and security auditing:
- CryptoDeepTools : A set of scripts for detailed cryptanalysis of the Bitcoin blockchain. This is a more general platform that combines various vulnerability research utilities. -35
- BTCDetect : A tool that implements the RingSide Replay Attack technique for practical private key recovery. Its architecture includes modules for entropy analysis, BIP-39/BIP-32 sequence reconstruction, and signature verification. -35
- KeyFuzzMaster : A fuzzing engine that systematically searches for weaknesses in signature verification code and elliptic curve operations. It is particularly effective at detecting wallets generated using 32-bit PRNG entropy, which greatly reduces the search space. -35
📚 Publications and materials of the center
The Center and its affiliated researchers publish their work on a variety of platforms. Key publications identified in the search:
- The Medium article , “Phantom Curve Attack: How the Pollard–Kangaroo Algorithm Exposes Private Keys of Lost Bitcoin…” details how KeySilentLeak collects signatures, analyzes entropy, and uses discrete logarithm solvers to recover keys. -21
- Articles on POLYNONCE : The main article about KeySilentLeak, as well as the related material “KeySilentLeak: A Cryptanalytic Tool for Recovering Bitcoin Wallet Private Keys.” -22
- Forum and blog discussions : Posts on Pikabu and forum.bits.media that duplicate and discuss materials about the development of KeySilentLeak and the scientific methods behind it. -23 -24
- Whitepapers at cryptodeeptech.ru : Research on specific cryptanalytic attacks, such as “Deep Cryptanalysis: Pohlig-Hellman Attack…” and “Deep Dive: GLV-Accelerated Pollard Rho Algorithm…”. -30 -31
⚙️ Vulnerability Analysis Methods
The center’s activities and the description of KeySilentLeak demonstrate a comprehensive methodology that combines theoretical cryptanalysis with practical implementation:
- Signature Collection and Statistical Analysis : The tool collects ECDSA signatures from the Bitcoin blockchain, extracting components
(r, s)and transaction hashes. It then performs statistical analysis to identify anomalies, such as duplicate valuesror correlations between parameters. -22 - Entropy analysis : Information-theoretical metrics are used to assess the quality of random number generation (nonce). Entropy is calculated using Shannon’s formula to identify low-entropy sequences. The Hamming weight distribution and chi-square tests are also used . -22 -23
- Solution of the Discrete Logarithm Problem (ECDLP) : Optimized algorithms are used to recover nonce
kfrom a public value :R = k⋅G- Pollard’s Kangaroo Algorithm : Used for interval search, reducing the search space from a theoretical 2²⁵⁶ operations to a feasible 2³² on modern GPUs. -21 -22
- Baby-step giant-step method : Used for small intervals. -22
- Pollard’s Rho algorithm : Used for the general case of the discrete logarithm. A GLV-accelerated modification for the secp256k1 curve is also explored . -22-31
- Private Key Recovery and Verification : After recovery,
kthe private keydis calculated using the formulad ≡ r⁻¹(s⋅k − H(m))(mod n). The result is verified by recalculating the public key and Bitcoin address. -22
Overall, the work of the Günther Zöeir Research Center exemplifies a deep academic and practical approach to the analysis of cryptographic vulnerabilities, combining theoretical research with the development of tools for their exploitation and auditing.
The Günther Zöeir Research Center is a research group specializing in cryptanalysis and blockchain security (specifically, Bitcoin). It is led by cryptographer and security specialist Günther Zöeir. privkeyroot.ru keysilentleak.ru
What does the center do?
Main areas of work:
- advanced cryptanalysis of elliptic curve cryptography (including the ECDSA algorithm – Elliptic Curve Digital Signature Algorithm);
- Bitcoin protocol security analysis;
- hardware wallet firmware audit;
- Cryptocurrency forensics and fund recovery;
- development of cryptanalytic tools for academic research.
One of the well-known tools created at the center is KeySilentLeak . It is a comprehensive system that analyzes and exploits vulnerabilities in Bitcoin cryptography implementations. It focuses on issues such as nonce reuse (special values used only once) and low-entropy nonce generation. Such errors can catastrophically reduce wallet security—theoretically, 256-bit security becomes virtually unbreakable in such cases. In their research, the group demonstrated, for example, the recovery of a private key for a wallet holding 60.7 BTC (approximately $7.6 million) by analyzing weak nonces generated due to an ESP32 hardware vulnerability (CVE-2025-27840). keysilentleak.ru cryptodeeptool.ru
Other tools
In addition to KeySilentLeak , the team is developing:
- CryptoDeepTools is a set of scripts for detailed cryptanalysis of the Bitcoin blockchain. It is a more general platform that brings together various vulnerability research utilities.
- BTCDetect is a tool that implements the RingSide Replay Attack technique for practical private key recovery. Its architecture includes modules for entropy analysis, BIP-39/BIP-32 sequence reconstruction, and signature verification
.
- KeyFuzzMaster is a fuzzing engine (vulnerability testing). It systematically searches for weaknesses in signature verification code and elliptic curve operations. It is particularly effective at detecting wallets generated using 32-bit PRNG entropy, as this greatly reduces the search space.
Publications
The Center does not publish classic scientific articles in peer-reviewed journals, but actively shares research results in other formats:
- Technical reports and blogs at zoeir.com.
2
- Video analysis and demonstrations on the YouTube channel @zoeirr.
3
- Materials are openly available on the ResearchGate platform, where researchers often post preprints and detailed analyses of specific cases.
6
I recommend starting with YouTube and the center’s website: they have many visual analyses of real attacks and explanations of how the tools work.
Vulnerability Analysis Methods
The team emphasizes a combination of classical and advanced mathematical approaches:
- Shannon entropy analysis is an assessment of the “randomness” of data to identify sources with low entropy (e.g., weak random number generators).
- Hamming weight histograms help detect patterns in the bit representations of data that may indicate vulnerabilities.
- The chi-square test is a statistical test for testing hypotheses about the distribution of data.
- Validation of pseudorandom number generators (PRNGs) against standards such as NIST FIPS 140-2.
1
- Nonce reuse attacks in ECDSA – if the same nonce appears in two signatures, this allows the nonce to be algebraically recovered and the private key to be calculated.
- Lattice techniques —for example, the LLL (Lenstra–Lenstra–Lovász) algorithm for solving problems like the Hidden Number Problem—allow keys to be recovered even when only a partial nonce is leaked (for example, if only a few bits of many signatures are known ).
- Side-channel analysis is the study of information leaks through the execution time of operations, energy consumption, or other physical characteristics, and not just through the algorithm itself.
1
- Space search algorithms —for example, the Pollard–Kangaroo algorithm—which significantly reduces the search space for key recovery, making the task feasible on modern GPU hardware.
1
In the Günther Zöeir Research Center methodology, private key recovery and verification is a strictly formalized process that can be roughly divided into two interrelated stages. The first stage is the calculation of the key d after the nonce has been recovered k . The second stage is cryptographic verification , confirming that the discovered key actually corresponds to the target Bitcoin address. The entire procedure is based on a fundamental vulnerability of ECDSA related to the quality of generation k.
🧮 Mathematical Foundation: Why It Works
The process relies on the basic ECDSA equation, which relates all the components of a signature:
s = k⁻¹ (H(m) + r ⋅ d) mod n
Where:
k— a secret one-time number (nonce).sandrare signature components.H(m)— hash of the signed message (transaction).d— long-term private key.n— the order of the group of points of the elliptic curve secp256k1.
The key point: if an attacker manages to discover k a signature, the private key d can be calculated using a simple algebraic permutation. The recovery formula is as follows:
d = r⁻¹ (s ⋅ k − H(m)) mod n -3 -15
Thus, the entire complexity of the attack comes down to finding k. For this purpose, the Günther Zöeir Research Center uses entropy analysis and discrete logarithm problem (ECDLP) solvers, such as the Pollard–Kangaroo algorithm, to find a weak value k from the recovered public value R = k ⋅ G. -1
🔬 Step-by-step recovery and verification process
The methodology implemented in the KeySilentLeak tool follows a clear algorithm.
Step 1: Private Key Recovery Module
This module comes into play after the discrete logarithm module has provided a candidate for the role of k. -3
- Calculation
d: For each signature for which was foundk, a calculation is performed using the formulad = r⁻¹ (s ⋅ k − H(m)) mod n. This calculation can be performed for either a single equation or a system of equations, if there are multiple signatures with knownk, to increase reliability. - Formation of candidates : The output is one or more candidates for the role of a private key.
Stage 2: Cryptographic Verification (Verification & Transformation Module)
This step is critically important as it eliminates false positives. The verification process includes the following steps: -1 -3
- Recalculating the public key : Given the computed private key
dand the base pointGof the secp256k1 curve, the tool computes the corresponding public pointQ' = d ⋅ G. - Comparison with the original : The resulting public key
Q'is compared with the public key extracted from the blockchain for the target address. If the match is found, the candidatedis considered correct. - Bitcoin Address Recalculation : For additional assurance, a Bitcoin address is regenerated from the public key
Q'(using the appropriate hashing). If this address matches the original target address, verification is considered fully successful. - Format conversion : After successful verification, the module can convert the found private key
dinto various formats, such as WIF (Wallet Import Format) , for easy import into wallets. -3
💡 Further information: Context and methods
Understanding this process would be incomplete without describing the methods that precede key recovery and the context in which they are applied.
Methods for finding weaknesses k
Recovery d is the final step. It’s preceded by the complex work of identifying the vulnerable k. The center uses a combination of methods:
- Statistical and entropy analysis : The tool looks for anomalies in signatures, such as repetition of values
r(indicating reusek) or statistical outliers indicating low nonce entropy. -3 - ECDLP (Discrete Logarithm Problem) Solution : Optimized algorithms are used to recover
kfrom a public value :R = k ⋅ G- Pollard–Kangaroo Algorithm : The basic method for searching
kin a known or limited range. It reduces the complexity from a theoretical 2²⁵⁶ to a practical 2³² operations on modern GPUs. -1 -2 - Other algorithms : The arsenal also includes the “baby-step giant-step” method for small intervals and Pollard’s “Rho” algorithm for the general case. -3
- Pollard–Kangaroo Algorithm : The basic method for searching
Attack Context: Dark Skippy and Phantom Curve Vulnerabilities
The center’s methodology is particularly effective against Dark Skippy attacks (also known as Phantom Curve Attacks ). This attack involves malware or hardware wallet firmware intentionally generating a k low-entropy nonce by directly deriving it from the user’s seed phrase. -23
For example, when signing two transactions, the first 8 bytes of the seed phrase can be used for k the first signature, and the second 8 bytes for the second. Thus, the two signatures in the blockchain implicitly encode the entire 16-byte entropy of the seed phrase. An attacker only needs to intercept these two signatures to recover , using the Pollard-Kangaroo algorithm k, and then the private key d using the formula above. -23
The Günther Zöeir Research Center’s private key recovery and verification methodology is a complete, scientifically proven pipeline. It begins with the collection and analysis of blockchain data, progresses through cryptanalysis (searching for weak keys k), then through algebraic key recovery d , and culminates in rigorous multi-stage verification that guarantees the accuracy of the result. This approach combines a deep understanding of the theoretical foundations of ECDSA with the practical implementation of computationally efficient algorithms.
Entropy analysis acts as a scout that finds anomalies and narrows the search area, and the Pollard–Kangaroo algorithm is a computational “weapon” that purposefully extracts the secret parameter k (nonce) from this narrowed area.
🔬 Implementation details of the Pollard–Kangaroo algorithm
In the KeySilentLeak architecture, this algorithm is implemented in the Discrete Logarithm Computation Module . Its task is to solve the elliptic curve discrete logarithm problem (ECDLP): find k from the equation R = k · G, subject to the condition that λ k is in a known or bounded interval [α, β]. -2
Step-by-step algorithm
The process described in the technical documentation of the center includes the following stages: -2
- Initialization and table sizing : The interval size is calculated
n = β - α. Based on this, the table size is determinedmas⌈√n⌉. This is a key parameter that balances memory usage and computation time. - Iteration function selection : A pseudorandom function is defined
f(X) = X + t_{Hash(X) mod k}. This function assigns a “jump” to each kangaroo based on the hash of the current point, ensuring pseudorandom but deterministic movement along the curve. - Construction of a Tame Kangaroos table : A “herd” of kangaroos is created, each starting from a known point
T₀ = α · G. Each kangaroo makes a series of jumpsTᵢ₊₁ = f(Tᵢ). All visited pointsTᵢand the “distance” (number of jumps) traveled to them are stored in a table. This table serves as a basis for comparison. - Wild Kangaroo Launch : One or more wild kangaroos are launched, starting from a target point
W₀ = R. They jump according to the same rulesf. - Collision detection : The “wild” kangaroos continue to move until one of their visited points
Wᵢmatches a pointT_jfrom a pre-built table of “tame” kangaroos. This event is called a collision. - Calculation
k: When a collision is detected (Wᵢ = T_j), the required noncekis calculated using a simple formula:k = α + (distance_j - distance_i), wheredistance_jis the distance traveled by the “tame” kangaroo to the collision point, anddistance_iis the distance traveled by the “wild” kangaroo. - Verification : The received value
kis immediately verified: it is calculatedk · Gand compared with the original targetR. A match confirms that the nonce was found correctly.
Key features of the implementation
- Complexity : The algorithm has a complexity of O(√n) operations on curve points. For a nonce with 64-bit entropy (
n ≈ 2⁶⁴), this is approximately 2³² operations , making the attack computationally feasible. For comparison, the theoretical complexity of a complete brute-force attack on a 256-bit key is 2²⁵⁶ operations. -2 -15 - GPU acceleration : KeySilentLeak’s implementation is optimized for modern graphics processing units (GPUs). The documentation provides a practical example: key recovery, which would take approximately 3,426 hours on a CPU (Intel Xeon E5-2680) , is completed in just 18.7 minutes on a GPU (NVIDIA RTX 4090) .
- Memory and Distributedness : There are variations of the algorithm that use kangaroo herds to increase the probability of collisions and that use distinguished points to reduce memory requirements, allowing for efficient parallelization of the computations.
📊 Methods of entropy analysis
This step is performed by the Entropy Analysis Module and is designed to quantitatively and qualitatively assess the “randomness” of the nonce parameters used in the collected signatures. -2
Information-theoretical metrics
The main metric is the Shannon entropy , which is calculated using the formula: -2
H = -∑ P(kᵢ) log₂ P(kᵢ)
where P(kᵢ) is the probability of using a specific nonce value kᵢ. In practice, for weak nonces with a known range of possible values, the entropy is calculated more simply:
H_weak = log₂(interval size)
For example, if a nonce is generated using only 8 bytes (64 bits), as in some Dark Skippy attack scenarios, its entropy is 64 bits instead of the required 256 bits . This 192-bit reduction in entropy makes the Pollard–Kangaroo algorithm feasible .
Statistical and heuristic tests
In addition to directly calculating entropy, the module applies a set of statistical tests to identify hidden patterns in the distribution of nonce bits: -2 -22
- Hamming Weight Distribution : The number of one bits in nonce values is analyzed. In a perfectly random sequence, the Hamming weight distribution should be close to binomial. Deviations may indicate bias in the generator.
- Chi-Square Test : This tests the hypothesis that the distribution of nonce values or its bits is uniform. The test quantifies how much the observed distribution differs from the expected uniform distribution.
- Signature-level heuristic checks : Checks for specific patterns that may indicate weakness: -2
- Duplicate values
r: If the same value appears in two different signatures of the same addressr, this is a clear sign of reuse of the same noncek. - Correlation between
rand other parameters : Analyzes whether there is a statistical relationship between the valuerand, for example, the message hashH(m)or other components of the signature. - Patterns in bit distribution
rands: Finding anomalies in the distribution of individual bits in signature components.
- Duplicate values
These methods allow not only to quantify entropy, but also to classify the type of nonce vulnerability, which determines the further strategy for its recovery.
