ZK Hints
Daily Digest 每日摘要 View All 查看全部
Lehmann et al. proposed a modular framework for anonymous credential systems, using core components and pluggable modules to support pseudonyms, range proofs, and device binding, improving standardization and security. Lehmann等人在论文中提出了一种匿名凭证系统的模块化框架，使用核心组件和可插拔模块支持伪名、范围证明和设备绑定等功能，以提升标准化和安全性。
Key Points: 要点：
- • Modular framework reduces redundant work, improving standardization of anonymous credential systems 模块化框架可减少重复工作，提升匿名凭证系统的标准化
- • Core components include commitments, signatures, and NIZK schemes, supporting pluggable extensions 核心组件包括承诺、签名和NIZK方案，支持可插拔扩展
- • Modules such as pseudonyms, range proofs, and device binding are independent of input, enhancing flexibility 模块如伪名、范围证明和设备绑定独立于输入，增强灵活性
- • Implementation using BBS signatures and ECDH, solving legacy hardware binding issues 实例化使用BBS签名和ECDSA，解决遗留硬件绑定问题
- • Framework suitable for digital identity and privacy pass applications 框架适用于数字身份和Privacy Pass等现实应用
- • Emphasizing security, modular design reduces protocol modification risks 强调安全性，模块化设计可降低协议修改风险
Jiang et al. proposed CoBBl, a SNARK compiler that combines direct translation and CPU simulation for dynamic constraint generation, significantly improving compilation and proving efficiency. Jiang等人在论文中提出了一种SNARK编译器CoBBl，结合直接翻译和CPU仿真的优势，实现动态约束生成，显著提升编译和证明效率。
Key Points: 要点：
- • CoBBl combines the advantages of direct translation and CPU simulation for dynamic constraint generation CoBBl结合直接翻译和CPU仿真优势，优化约束生成
- • Significantly improves compilation and proving efficiency 性能显著提升：编译时间比CirC快1-30倍，证明时间快26-350倍
- • Avoids the cost of unexecuted calculations, improving efficiency 在Jolt友好基准上，证明时间优于Jolt 1.1-1.8倍
- • Supports program-specific optimization, reducing constraint count 避免未执行计算成本，提高效率
- • 支持程序特定优化，减少约束数量
@ziskvm team announced a major proving milestone for ZisK, with all Ethereum blocks now provable in real-time, with an average proving time of about 7.5 seconds, improving ZK proof efficiency. @ziskvm 团队宣布ZisK实现重大证明里程碑，所有以太坊区块现可实时证明，平均证明时间约7.5秒，提升了ZK证明效率。
Phantom VM team shared a blog post about FHE-RAM implementation, based on Poulpy library, supporting private read and write of encrypted RAM, with server knowing only RAM size and operation type, ensuring data privacy. Phantom VM 团队在博客中分享了FHE-RAM的实现，基于Poulpy库，支持加密RAM的私有读写，服务器仅知RAM大小和操作类型，确保数据隐私。
Key Points: 要点：
- • FHE-RAM allows encrypted storage of RAM, with private read and write by users, and server knowing only operation type and RAM size FHE-RAM允许加密存储RAM，用户可私有读写，服务器仅知操作类型和RAM大小
- • GLWE and GGSW cryptography techniques support address decomposition and homomorphic operations 使用GLWE和GGSW密码学技术，支持地址分解和同态操作
- • 1MB RAM read delay is 302ms, write delay is 871ms, with parallel optimization potential 1MB RAM在单核i9-12900K上读延迟302ms，写延迟871ms，可并行优化
- • Design includes key generation, data encryption, and read-write protocol, ensuring end-to-end privacy 设计包括密钥生成、数据加密和读写协议，确保端到端隐私
- • Open-source implementation provides detailed parameters and examples, facilitating integration and extension 开源实现提供详细参数和示例，便于集成和扩展
- • Suitable for distributed computing scenarios requiring high privacy protection 适用于需要高隐私保护的分布式计算场景
Stanford Blockchain Club shared a blog post with Dan Boneh, discussing LatticeFold's post-quantum security, hardware friendliness, and folding techniques, as well as its applications in ZK proof systems. Stanford Blockchain Club在博客中分享了与Dan Boneh的对话，讨论了LatticeFold架构家族及其在ZK证明系统中的应用，包括后量子安全、硬件友好算术和折叠技术等内容。
Key Points: 要点：
- • LatticeFold uses Ajtai commitments and MSIS assumption, providing post-quantum security, replacing traditional discrete logarithm commitments LatticeFold使用Ajtai承诺和MSIS假设，提供后量子安全，替代传统离散对数承诺
- • Folding technology controls norm growth through extension, decomposition, and folding steps, ensuring binding security 折叠技术通过扩展、分解和折叠步骤，控制范数增长，确保绑定安全
- • Supports 64-bit arithmetic, friendly with CPU/GPU hardware, improving proof efficiency 支持64位模数算术，与CPU/GPU硬件友好，提升证明效率
- • Can be applied to aggregated post-quantum signatures, enhancing blockchain system scalability 可应用于聚合后量子签名，增强区块链系统的可扩展性
- • Compared to existing schemes like HyperNova, LatticeFold has advantages in hardware implementation and security 与HyperNova等现有方案相比，LatticeFold在硬件实现和安全性上有优势
- • Emphasizes the complementary nature of folding and ZK, folding handles scalability, ZK handles privacy 强调折叠与ZK的互补性，折叠处理可扩展性，ZK处理隐私
Pratyush Mishra organized CIS 7000 course, running on Mondays and Wednesdays in Fall 2025, focusing on the mathematical foundations, implementation, and applications of zkSNARKs. Pratyush Mishra组织了CIS 7000课程，时间为2025年秋季的周一和周三，从8月底开始到12月中旬结束，专注于zkSNARKs的数学基础、实现和应用。
Key Points: 要点：
- • The course does not have public recordings, but the slides are available. 课程无公开录屏，但可查看幻灯片。
- • Course covers mathematical foundations, implementation, and applications of zkSNARKs 课程涵盖zkSNARKs的数学基础、实现和应用
- • Includes key technologies such as multi-variate IOPs, KZG commitments, and Spartan PIOP 包括多项式IOPs、KZG承诺、Spartan PIOP等关键技术
- • Emphasizes practical applications of zkSNARKs in blockchain and privacy protection 强调zkSNARKs在区块链和隐私保护中的实际应用
- • Students deepen understanding of ZKP systems through paper presentations and projects 学生通过论文展示和项目深入理解ZKP系统
- • Course structure combines lectures, discussions, and experiments to enhance comprehensive abilities 课程结构结合讲座、讨论和实验，提升综合能力
- • Final project can involve research or literature review, driving academic progress
Binyi Chen et al. in paper proposed Symphony, a lattice-based high-arity folding SNARK, avoiding hash embedding in SNARK circuits, with memory efficiency, parallelization, and post-quantum security. Binyi Chen等人在论文中提出了Symphony，一种基于格的高阶折叠SNARK，避免在电路中嵌入哈希函数，具有内存高效、并行化和后量子安全等特性。
Key Points: 要点：
- • Symphony avoids hash embedding in SNARK circuits, reducing security risks and proof overhead Symphony避免在SNARK电路中嵌入哈希函数，减少安全风险和证明开销
- • Adopting lattice-based folding scheme, providing post-quantum security guarantees 采用格基折叠方案，提供后量子安全保证
- • Supporting memory efficiency, parallelization, and streaming processing, optimizing performance 支持内存高效、并行化和流式处理，优化性能
- • Proof size and verification time are multi-polynomial, enhancing scalability 证明大小和验证时间为多对数级，提升可扩展性
- • Universal compiler can convert folding scheme to SNARK, enhancing flexibility 通用编译器可将折叠方案转换为SNARK，增强灵活性
- • Having potential advantages in zkVM and proof learning applications 在zkVM和证明学习等应用中具有潜在优势
@barrywhitehat in post discussed how to implement private EVM by adding pstore and pload opcodes, supporting private user states, analyzing the trade-off between privacy and global state. @barrywhitehat 在帖子中讨论了如何通过添加pstore和pload操作码实现私有EVM，支持私有用户状态，同时分析了隐私与全局状态的权衡。
Key Points: 要点：
- • Private global state is infeasible, as the prover must know the state to generate proofs 私有全局状态不可行，因证明者需知晓状态才能生成证明
- • Adding pstore and pload opcodes supports private user states 添加pstore和pload操作码可支持私有用户状态
- • Private storage trees and invalid trees manage private data 使用私有存储树和无效树来管理隐私数据
- • Need to handle metadata leaks, such as message.sender and tx.origin 需处理元数据泄露问题，如message.sender和tx.origin
- • Can achieve private rollup, with users or servers generating proofs 可实现私有rollup，用户或服务器生成证明
- • Future needs to consider EOA privacy and dynamic value storage race conditions 未来需考虑EOA隐私和动态值存储的竞态条件
Zyskind et al. in paper proposed an efficient threshold FHE decryption protocol, avoiding noise proliferation, through secure MPC rounding to achieve simulation security in the UC framework, with online throughput improvement of approximately 20,000 times. Zyskind等人在论文中提出了一种高效阈值FHE解密协议，避免噪声泛滥，通过安全MPC舍入实现UC框架下的模拟安全，在线阶段吞吐量提升约20,000倍。
Key Points: 要点：
- • Protocol avoids noise proliferation, reducing overhead and parameter size 协议避免噪声泛滥，减少开销和参数大小
- • Offline/online design, preprocessing reduces online communication and computation cost 采用离线/在线设计，预处理降低在线通信和计算成本
- • Supports multiple adversary models, such as dishonest majority or honest majority 支持多种敌手模型，如不诚实多数或诚实多数
- • Online throughput improvement of approximately 20,000 times, with a latency optimization of up to 37 times 在线阶段吞吐量提升约20,000倍，延迟优化达37倍
- • Technique can be applied to other MPC protocols, such as secure machine learning 技术可应用于其他MPC协议，如安全机器学习
- • Proven secure in the UC framework, enhancing composability and practicality 在UC框架下证明安全，增强可组合性和实用性
@VitalikButerin in blog post detailedly explained the GKR protocol, including its application in ZK proof systems, the principle of Sumcheck protocol, and the proof optimization of Poseidon hash function. @VitalikButerin 在博客中详细讲解了GKR协议，包括其在ZK证明系统中的应用、Sumcheck协议原理以及Poseidon哈希的证明优化等内容。
Key Points: 要点：
- • GKR for batched layered computations, no intermediate commitments GKR为批量分层计算设计，无需中间层承诺
- • Uses sumcheck to recursively transform verification 基于Sumcheck协议，递归转化验证问题
- • Specialized for hashes and neural networks 专攻哈希和神经网络等结构化计算
- • Optimized overhead <15x theoretically, single-digit practically 优化后理论开销<15x，实践可达个位数
- • Requires ZK-SNARK/STARK for privacy 需结合其他系统实现零知识
- • Extensions enable near-zero overhead at scale 支持扩展优化，宽度增加时开销趋零
@brevis_zk announced Pico Prism (zkVM), which achieved significant performance improvements: 98.9% transaction coverage within 10 seconds, hardware cost reduced by 50% to $128,000, proof time shortened to about 6 seconds. The scheme aims to break through Ethereum verification bottlenecks, with all benchmark open-sourced. @brevis_zk 推出Pico Prism (zkVM)，性能大幅提升：10秒内交易覆盖率升至98.9%，硬件成本降50%至12.8万美元，证明时间缩短至约6秒。该方案旨在突破以太坊验证瓶颈，所有测试已开源。
Grassi et al. in paper proposed Poseidon2b, a hash function designed for binary extension fields, inheriting the circuit-friendly characteristics of Poseidon2, and compatible with Binius et al. proof systems, analyzing algebraic cryptanalysis and subspace trajectory attacks. Grassi等人在论文中提出了Poseidon2b，这是一种专为二进制扩展字段设计的哈希函数，旨在继承Poseidon2的电路友好特性，并与Binius等证明系统兼容，同时分析了代数密码分析和子空间轨迹等攻击向量。
Key Points: 要点：
- • Poseidon2b is designed for binary extension fields, inheriting the circuit-friendly characteristics Poseidon2b专为二进制扩展字段设计，继承电路友好特性
- • Compatible with Binius et al. proof systems, optimizing proof and verification time 与Binius等证明系统兼容，优化证明和验证时间
- • Re-evaluating Poseidon and Poseidon2 attacks in binary extension fields 重新评估Poseidon和Poseidon2的攻击在二进制字段的适用性
- • Focusing on algebraic cryptanalysis and subspace trajectory attacks 重点关注代数密码分析和子空间轨迹等攻击向量
- • Benchmarking shows advantages in proof size, time, and verification 基准测试显示在证明大小、时间和验证方面有优势
- • Security analysis highlights the unique attack risks of binary extension fields 安全性分析强调二进制字段特有的攻击风险
Nick et al. in paper proposed Shielded CSV protocol, combining PCD abstraction to achieve efficient privacy transactions, requiring only 64 bytes of data written to the blockchain, improving Bitcoin privacy to 100 transactions per second. Nick等人在论文中提出Shielded CSV协议，结合PCD抽象实现高效隐私交易，仅需64字节数据写入区块链，提升比特币隐私至每秒100笔交易。
Key Points: 要点：
- • Shielded CSV provides the first truly private transaction construction, requiring only 64 bytes of nullifier written to the blockchain Shielded CSV提供首个真正私有交易构造，仅需64字节nullifier写入区块链
- • Verification cost is low, users only need a single Schnorr signature verification, non-users can ignore data 验证成本低，用户仅需单个Schnorr签名验证，非用户可忽略数据
- • Proof size is independent of transaction history, improving efficiency and privacy 证明大小独立于交易历史，提升效率和隐私
- • Based on PCD abstraction, supports folding schemes and recursive STARKs etc. implementation strategies 基于PCD抽象，支持折叠方案和递归STARKs等实现策略
- • Extremely scalable, future can further improve the framework and protocol 可扩展性强，未来可进一步改进框架和协议
- • High security, protected by ZK proofs, reducing data leakage risk 安全性高，通过ZK证明保护交易隐私，减少数据泄露风险
Zhang et al. in paper proposed a dynamic zkSNARK scheme DYNARK, which can update proofs in sublinear time when witnessing small changes, based on Groth16 protocol to optimize efficiency. Zhang等人在论文中提出了一种动态zkSNARK方案DYNARK，可在见证小变化时以次线性时间更新证明，基于Groth16协议优化效率。
Key Points: 要点：
- • DYNARK supports fast proof updates when witnessing small changes, reducing computational overhead DYNARK支持见证小变化时的快速证明更新，减少计算开销
- • Update time is O(d) group operations in semi-dynamic settings, O(d√n log n) in fully dynamic settings 半动态设置下更新时间为O(d)群操作，全动态为O(d√n log n)
- • Proof size remains 192 bytes, verification time 4.4 milliseconds, consistent with Groth16 证明大小保持192字节，验证时间4.4毫秒，与Groth16一致
- • Experiments show that with n=2^20, preprocessing takes 74.3 seconds and updates take 3-60 milliseconds 实验表明，n=2^20时，预处理74.3秒后更新仅需3-60毫秒
- • Does not depend on pairing product parameters or other zkSNARKs, simplifying system design 无需依赖配对乘积参数或其他zkSNARK，简化系统设计
- • Fully compatible with existing Groth16 infrastructure, making deployment easier 完全兼容现有Groth16基础设施，便于实际部署
Hafezi et al. in paper systematize the design of lookup table arguments, propose a unified framework covering various lookup types, and evaluate existing protocols in terms of proof cost, composability, etc. Hafezi等人在论文中系统化分析了查找表论证的设计，提出统一框架覆盖多种查找类型，并评估现有协议在证明成本、可组合性等方面的表现。
Key Points: 要点：
- • Lookup table arguments are widely used in non-interactive operations, zk-VM and set membership proofs 查找表论证广泛用于非本地操作、zk-VM和集合成员证明
- • Existing constructions differ in assumptions, efficiency, and composability 现有构造在假设、效率和可组合性上差异显著
- • Introduce a unified framework covering standard, projection, index, vector and decomposable lookups 引入统一框架覆盖标准、投影、索引、向量和可分解查找
- • Classify protocols based on proof technology and composition style 分类协议基于证明技术和组合风格
- • Evaluation metrics include prover cost, table size dependence, and recursive proof composability 评估维度包括证明者成本、表大小依赖和递归证明兼容性
- • Provide practical guidance on selection and highlight the advantages and limitations of preprocessing and decomposability 提供实践选择指南并突出预处理和可分解性的优势与限制
@0xLita team released Valida zkVM 1.0, including parallel proving and compiler optimizations, improving performance and developer experience, but emphasizing that it is not yet production-ready. @0xLita 团队发布了Valida zkVM 1.0，包括并行证明和编译器优化等特性，提升了性能和开发者体验，但强调其尚未达到生产就绪状态。
Key Points: 要点：
- • Valida zkVM 1.0 introduces parallel proving, significantly accelerating multi-threaded workloads Valida zkVM 1.0引入并行证明，可显著加速多段工作负载
- • Compiler toolchain optimizations include smart branching and operation folding, improving code efficiency 编译器工具链优化包括智能分支和操作数折叠，提升代码效率
- • New Boolean constraints and program consistency checks strengthen proof correctness 新增布尔约束和程序一致性检查，强化证明正确性
- • Security issue: Soundness auditing is still in progress, not recommended for production environment 安全性问题：健全性审查仍在进行，不建议用于生产环境
- • Thanks to Hideaki Takahashi for fixing critical soundness errors, promoting collaborative development 感谢Hideaki Takahashi修复关键健全性错误，推动协作开发
- • Future plans include expanding performance, language support, and application scale 未来计划扩展性能、语言支持和应用规模
@PrivacyEthereum team released September 2025 newsletter, covering ZK-Kit progress and PSE roadmap, focusing on the application of zero-knowledge technology in decentralization and privacy protection. @PrivacyEthereum 团队发布了九月通讯，涵盖ZK-Kit进展和PSE路线图，聚焦零知识技术在去中心化和隐私保护方面的应用。
Key Points: 要点：
- • The ZK-Kit team focuses on developing zero-knowledge tools, enhancing programming cryptographic capabilities ZK-Kit团队专注于零知识工具开发，提升编程密码学能力
- • PSE roadmap outlines 2025 and future strategic directions in the zero-knowledge domain PSE路线图概述2025年及未来在零知识领域的战略方向
- • Emphasizing decentralization and privacy protection in zero-knowledge technology 强调去中心化和隐私保护在零知识技术中的重要性
- • Newsletter content covers open-source project updates and community dynamics 通讯内容涵盖开源项目更新和社区动态
- • Zero-knowledge proof technology is driving blockchain privacy and scalability development 零知识证明技术正推动区块链隐私和可扩展性发展
- • The team continues to optimize tools to support developers in building ZK applications 团队持续优化工具以支持开发者构建ZK应用
Researchers independently published two papers disclosing physical attacks Battering RAM and Wiretap against Intel SGX and AMD SEV-SNP, exploiting deterministic encryption vulnerabilities, allowing attackers to view and manipulate protected data. 研究人员独立发布两篇论文，披露针对Intel SGX和AMD SEV-SNP的物理攻击Battering RAM和Wiretap，利用确定性加密漏洞，允许攻击者查看和操纵受保护数据。
Key Points: 要点：
- • TEE depends on deterministic encryption, vulnerable to replay attacks, requiring hardware changes to enhance security TEE依赖确定性加密，易受重放攻击，需改硬件以增强安全
- • Battering RAM attack costs less than $50, supporting active data read and write Battering RAM攻击成本低于50美元，支持数据读写和篡改
- • Wiretap attack costs $500-1000, supporting passive data read only Wiretap攻击成本500-1000美元，仅支持被动数据读取
- • Attacks through physical interceptors, supply chain or physical access can lead to vulnerability exploitation 攻击通过物理拦截器实现，物理访问可导致漏洞利用
- • Chip manufacturers claim TEE design does not defend against physical attacks, but cloud services still widely rely on it 芯片商称TEE设计不防御物理攻击，但云服务仍广泛依赖
- • Blockchain services like Phala use TEE to protect smart contracts, facing potential security risks 区块链服务如Phala使用TEE保护智能合约，面临潜在安全风险
@risczero team disclosed a high-risk vulnerability in the zkVM platform, involving a memory safety issue in the sys_read function, which could lead to arbitrary code execution. The vulnerability has been fixed. @risczero 团队在GitHub安全公告中披露了zkVM平台中的高危漏洞，涉及sys_read函数内存安全问题，可能导致任意代码执行，已发布修复版本。
Key Points: 要点：
- • Vulnerability allows host to write arbitrary code to guest memory via sys_read 漏洞允许主机通过sys_read写入guest任意内存位置，执行任意代码
- • All guest programs using affected versions are at risk, breaking zk guarantees 所有使用受影响版本的guest程序均存在风险，严重破坏ZK保证
- • Fix has been applied to risc0-zkvm versions 2.3.2 and 3.0.3, removing unsafe pointer operations 修复已应用于risc0-zkvm版本2.3.2和3.0.3，移除不安全指针运算
- • Developers need to update Cargo.toml with related crate versions and rebuild application 开发者需更新Cargo.toml中相关crate版本并重建应用
- • RISC Zero proof system and circuits are not affected, no prover action needed RISC Zero proof系统和电路未受影响，无需prover行动
- • Projects using risc0-aggregation or RiscZeroSetVerifier need to upgrade to >=0.9 version 使用risc0-aggregation或RiscZeroSetVerifier的项目需升级至>=0.9版本
@hexensio team compared Halo2, Zirgen, and Plonky3 three zkVM DSLs in their blog, analyzing their differences in circuit architecture, data model, and constraint expression. @hexensio 团队在博客中比较了Halo2、Zirgen和Plonky3三种zkVM DSL，分析了它们在电路架构、数据模型和约束表达方面的差异。
Key Points: 要点：
- • Three DSLs: Halo2 (PLONKish matrix), Zirgen (STARK trace), Plonky3 (AIR direct constraints) 三大DSL：Halo2（PLONKish矩阵）、Zirgen（STARK踪迹）、Plonky3（AIR直接约束）
- • Architecture: Matrix vs state transitions vs polynomial identities 架构差异：矩阵vs状态转移vs多项式恒等式
- • Data model: Typed columns vs register composition vs typeless columns 数据模型：类型化列vs寄存器组合vs无类型列
- • Public interface: Instance columns vs keyword vs explicit binding 公共接口：实例列vs关键字vs显式绑定
- • Constraint method: Rotation gates vs cross-cycle references vs guard mechanisms 约束方式：旋转门vs跨周期引用vs守卫机制
- • Core trade-off: Abstraction vs control vs development efficiency 核心权衡：抽象度vs控制力vs开发效率