Applicant Information
Name of Applicant & Applicant’s Representative [If Applicable]:
- Stephen Tong (CEO and Co-Founder, Zellic);
- Kaushik Swaminathan (Head of Growth, Zellic, Representative)
Email Address: stephen@zellic.io; kaushik@zellic.io
Telegram Handle (if applicable): k_zellic
LinkedIn Profile (if applicable): Zellic | LinkedIn
Role being applied for [1 Max]: Security-Oriented Member
Background Information
We are Zellic, a leading blockchain security firm. Security providers have historically sought large amounts of funding from DAOs and other public goods organizations for work of dubious value to token holders. Meanwhile, projects often spend a third of their budget on security. We believe this prescriptive, extractive approach is fundamentally misaligned with the long-term interests of the crypto community and Web3 industry.
Counter to this trend, we propose a radically-aligned “pay for what you get” model. In this model, we will only be compensated for the deliverables that token holders actually want and find useful. Meanwhile, we propose deliverables that directly minimize the amount of security spend needed by developers in the Arbitrum ecosystem to ship and deploy code.
Beyond fees and payment structure, we directly prove and enforce our alignment to Arbitrum token holders financially:
- Voluntary token bond. If Zellic is selected for the ARDC, we will at our own expense post a $75,000 USD bond to be locked up in staked ARB tokens during the performance of this proposal.
- Vesting. We also voluntarily request that all payments for services under this proposal vest over a predetermined period. These terms ensure that Zellic will remain fully committed and faithful to ARB token holders first and foremost.
About us
Zellic is a vulnerability research firm with deep expertise in blockchain security. We specialize in EVM, Move (Aptos and Sui), and Solana, as well as ZK and Cosmos. We identify complex vulnerabilities and prevent catastrophic security events.
Among others, LayerZero, StarkWare, SushiSwap, and the Solana Foundation trust Zellic to secure their future. We review L1s and L2s, cross-chain protocols, wallets and applied cryptography, web applications, and more. We also have a dedicated zero-knowledge cryptography team, and work closely with projects like Scroll, Axiom, and Succinct Labs.
Zellic is led by Stephen Tong and Jasraj Bedi, who previously founded the #1 CTF team worldwide in 2020 and 2021. Our engineers bring a rich set of skills and backgrounds, including cryptography, web security, mobile security, low-level exploitation, and finance. We’re also a founding member of the Security Alliance (SEAL) led by samczsun, an industry effort to raise the bar for blockchain security.
We believe that a competent security member of ARDC should support all subject-matter areas relevant to the DAO: Governance Proposals; Research; Framework Development; Risk Assessment; Secure Code Reviews; Threat Modeling; Testing Enhancements; and Retroactive Analysis Of Passed Initiatives.
Objectives and Motivations
We want to empower DAO members to make good decisions, especially when they involve security and/or technical evaluation. Our objective is to increase Arbitrum’s TVL and developer adoption. We will do this by providing useful deliverables that (1) make users more eager to use applications built on Arbitrum and (2) make it easier and cheaper for developers to build and ship those applications in the first place.
Arbitrum has almost $3B in TVL, but insufficient tools, processes, or advisory to ensure DAO governance and ecosystem development are secure. We want to bring in battle-tested best practices, tooling, and services to bolster the security posture of the Arbitrum ecosystem at-large. And most importantly, we want to do so in a way that is actually useful for token holders.
Skills and Experience
Below are Zellic’s areas of expertise that are most relevant to Arbitrum DAO:
Arbitrum Ecosystem: We have audited and/or have on retainer several leading protocols on Arbitrum, including Hyperliquid, Timeless, Perennial Finance, Y2K, and Premia.
L1s, L2s, and Roll-ups: We work with some of the largest L1s—Solana Foundation, Aptos Labs, and Mysten Labs—and L2s—StarkNet, Scroll, and Mantle—to identify bugs in networks, application layers, custom precompiles, and more. Earlier this year, we discovered a critical vulnerability in Move’s bytecode verifier that put billions of dollars at risk across multiple chains.
Cross-chain apps: We are experts in cross-chain infrastructure, working closely with bridges like LayerZero, Wormhole, Succinct Labs, and more. Our commitment to bridge security extends to governance support; we are a member of Uniswap’s Bridge Assessment Committee. In addition, we’ve reviewed cross-chain applications including Stargate, Tristero, Catalyst, Shrapnel, Cedro, and the OFT/ONFT standards.
Web application security: Zellic has deep expertise spanning the full stack of web applications spanning all the major tech stacks—ranging from dynamic backends built in Javascript, Python, and Ruby, to high-performance APIs in Golang and Rust, to client-side applications in React and Electron. Our team comes with extensive CTF and bug bounty experience, and has claimed bounties from the largest technology companies and public institutions including the US Department of Defense, Github, Yahoo, Shopify, PayPal, Adobe, CrowdStrike, Amazon, and Bitfinex. In 2018, our CTO Jasraj Bedi found a novel DNS rebinding bug in Geth—drawing on our capabilities in web and native security.
Applied cryptography: Zellic has a dedicated team of strong theoretical and applied cryptographers. Implementing cryptographic applications securely, like Web3 wallets, is incredibly difficult. We help clients navigate a minefield of potential pitfalls and mistakes. Wallets we’ve reviewed include Pontem, Avara (Aave Lens), Lootrush, and Ledger. We’ve also reviewed adaptors such as Aptos IdentityConnect. We helped Privy secure their Shamir’s Secret Sharing (SSS) implementation used by friendtech. We look at non-custodial wallets, ERC4337 (AA), MPC, SSS, EOAs, native multisig support, enclave solutions, and social login and key recovery.
Secure Enclaves, TEEs, and Trusted Computing: Zellic’s background in native application security is unparalleled: We love C, C++, x86_64, Aarch64, and SystemVerilog. We work with embedded firmware, secure elements, and hardware wallets. We worked with the Solana Mobile team to conduct a full-stack review of their secure enclave architecture from the Android app down to the TEE. We also worked with the Interchain Foundation to review their Cosmos Ledger integration, where we uncovered and fixed critical memory corruption vulnerabilities. We’re also experts with cloud enclave stacks like KMS and AWS Nitro Enclaves. Trusted applications demand high assurance and deep systems knowledge, which our clients trust us to deliver.
Formal verification, Static Analysis, and Fuzzing: Zellic are experts at program analysis, leveraging tools such as fuzzers, SMT solvers like CVC5 and Z3, LLVM, symbolic executors, and proof assistants like Lean and Coq. Our team includes researchers who have published peer-reviewed fuzzing papers and static analysis specialists. Using Z3, we formally proved the security of WETH, the world’s most relied-upon smart contract. With Pontem, Laminar Markets, and PancakeSwap, we leveraged the Move prover to guarantee critical protocol invariants. With Mysten Labs, we thoroughly fuzzed the Move VM for deep property-based testing. These techniques create additional assurance beyond a thorough manual review.
Zero-knowledge circuits: Zellic’s dedicated zero-knowledge team combines a distinguished skill set in advanced cryptography, vulnerability research, and competitive hacking. We review circuits in Circom and Halo2 for zkEVMs, zkVMs, privacy and identity protocols, and interoperability infrastructure. Our clients include rollups (Scroll), coprocessors (Axiom), privacy primitives (Nocturne), and zk-bridges (Polyhedra).
Proposal Review & Assistance
Our proposal below outlines a number of initiatives of varying complexity and application that we are confident to execute at the highest level. That being said, we do not want to be presumptuous or heavy-handed in what the community needs.
We will defer to governance participants to determine which of the initiatives is most appropriate for the community for the duration of our six-month tenure.
1. Template primitives for Arbitrum Stylus
We want to minimize audit spend for Arbitrum developers. By providing pre-audited primitives, Arbitrum developers will be able to fork and ship code quickly and safely.
Stylus has the potential to be a step-change in developer experience across all EVM chains. It allows developers to write smart contracts in programming languages that compile down to WASM, such as Rust, C, C++, and many others. Given its novelty, however, there are few resources or projects that demonstrate its robustness.
Zellic can build secure templatized primitives for Stylus–e.g., an ERC20 implementation–that can be leveraged to develop higher-order dApps like AMMs, perps DEXs, lending protocols, and more. These templates would be pre-audited by our security researchers. Our goal is to increase the baseline level of ecosystem security, even for permissionless deployments of long-tail assets and protocols.
2. Guidelines and frameworks for AIPs
AIPs are often well-intentioned but poorly drafted because there are no standard guidelines. Alongside other ARDC members, we will develop a set of standards and best practices for AIPs so that proposers and evaluators have the necessary information to make an informed decision. Specifically, a lot of proposals fail to account for overall ecosystem impact that even minor changes can have. This protects Arbitrum token value by eliminating security issues before they are even developed.
For every governance proposal that relates to our security expertise and/or proposes to spend DAO budgets on audits, we will contribute a forum post outlining security considerations for that project for the community’s benefit in decision-making. For instance, we will contribute input on ways proposed audit costs could be minimized.
We will also conduct full security reviews of any on-chain governance upgrades as needed, at the discounted rates outlined later in this proposal.
3. Arbitrum CTF competitions
Capture The Flag competitions are the epicenter of security research. They consist of a set of computer security puzzles involving reverse-engineering, memory corruption, cryptography, web technologies, and more. CTFs and similar code competitions are a verified go-to-market strategy to attract high-quality developers to a new ecosystem. Curta competitions, for instance, have discovered some of the best protocol engineers via their programming competition platform on EVM. We have designed several Curta challenges. Some other recent Web3 efforts of ours include MoveCTF, Ingonyama ZK CTF, and Paradigm CTF.
We’ve led the #1 ranked CTF team worldwide in 2020, 2021, and 2023, and have won some of the most prestigious competitions including GoogleCTF, Real World CTF, PlaidCTF, and DEF CON Quals. With the rapid growth of rollups and sovereign blockchains, Arbitrum-specific CTFs—which are operationally complex with high technical barriers to organize—will be a strategic way to identify and attract top developer talent to the ecosystem.
4. Security primers on Arbitrum
Zellic will write security primers, case studies, PSAs, and other analysis for the community’s benefit. We will write these in the style of our previous educational blog posts. Similar security primers that we’ve written for other ecosystems include: Aptos; Sui; Cairo; ZK. We expect that ongoing commitment to education and documentation will distinguish Arbitrum as a mature, robust platform.
In addition to our work for clients, Zellic closely follows all ongoing critical exploits and hacks in the crypto ecosystem. On multiple occasions, our security researchers have successfully reverse-engineered several major attacks as they were ongoing. For example, collaborating with samczsun, Zellic was the first team to triage and reverse engineer the $325,000,000 attack on the Wormhole bridge in February 2022.
During these times of crisis, our auditors regularly publish long-form Twitter threads to help raise the crypto ecosystem’s awareness and education regarding security. Public education and community engagement are important pillars of Zellic’s ethos. Below are select Twitter threads that received the highest user engagement and impressions this year, and we envision publishing similar threads and additional PSAs for Arbitrum DAO: Ledger wallet drain; Nomad $190M bridge hack; Slope wallets hack; meta-analysis of cross-chain bridge exploits.
5. Governance incident response
In addition to reviewing on-chain governance proposals, we will run governance attack simulations to ensure that key stakeholders within the DAO are well-prepared for emergency situations–low probability but critical impact incidents. Bad actors can pass malicious proposals, even in forums with a wide range of active participants. Processes around incident response will establish Arbitrum as a leader in ecosystem security.
In such scenarios, key governance delegates must move swiftly and with a clear operating procedure to reverse the malicious proposal and/or its impact. We will design an Arbitrum DAO-specific playbook for this. Our experience as a founding member of SEAL is testament to our commitment to incident preparedness and response.
6. Arbitrum Drift Tracker
Zellic has built a tool called the Audit Drift Tracker, which tracks what code is audited and unaudited for the biggest DeFi protocols. ‘Drift’ specifically refers to the difference between code that is audited and code that is deployed. To the best of our knowledge, no one is tracking the on-chain audit drift of popular DeFi and Web3 protocols. Audit drift was at the heart of the $190M Nomad bridge exploit.
Given the pace of development in the ecosystem, we want to devote resources for an exclusive platform to track audit drift in leading Arbitrum projects. TVL security is among the highest priorities for any chain, and Drift Tracker offers both developers and users an accessible tool to make security-informed decisions and hold protocols accountable. We expect this should have a meaningful impact on growing TVL within the ecosystem as users feel more equipped to better diligence their counterparties on-chain.
7. Forky (Forks Identifier)
Forky is a tool that presents the smart contract differences between a fork of a protocol and a base (parent) protocol. For example: PancakeSwap is a fork of Uniswap v3. This tool would allow you to easily view the differences between both codebases and their corresponding risk considerations. This will allow both retail traders and institutional funds to deploy with confidence on counterparty security.
DeFi protocols are particularly susceptible to fork-related exploits. Users often assume that forks carry the same security assumptions of its parent protocol, without checking and/or understanding the changes that can be made. Forky highlights in plain English the scope of changes as well as its intended effect, like changes in permissions, admin functions, use of standards, etc.
We will build an Arbitrum-focused Forky that allows users to input the source code of any new protocol on Arbitrum and compare it to the most-forked protocols on Ethereum–like Uniswap, Aave, Compound, etc.
8. Static Analyzer for Arbitrum Stylus
We can build a static analyzer for Arbitrum Stylus, similar to what Slither is for Solidity. However, we believe that the development cost here would be prohibitive, and this tool would ultimately not be directly value-accretive to ARB developers. While it is within our capabilities, as we have significant experience with compiler engineering and static analysis, we do not provide a cost estimate for this component. Should the DAO believe that this component would be useful for the Arbitrum community, we would be happy to amend our proposal to include this component.
Review on Chain Proposal Code Updates
We’re respected as security researchers because we find bugs that others miss. Listed below are some of our recent security findings in leading blockchain projects. We hope it demonstrates our expertise across the wide range of security considerations that should be top-of-mind for Arbitrum DAO.
Our commitment to security research includes governance support. We are a member of Uniswap’s Bridge Assessment Committee. In February 2023, the Uniswap Foundation convened this committee to evaluate cross-chain bridges in DAO governance. We evaluated six bridges and approved two for the DAO’s cross-chain governance use case, and determined that a multi-bridge architecture was likely the best option for Uniswap.
In our review of Cosmos SDK’s new string-based sign mode for signing with hardware devices, we identified a buffer overflow in the function responsible for converting ASCII control characters to their escape sequence equivalents. An attacker could have abused this bug to overwrite the stack of the ledger and partially control the instruction pointer and registers.
As part of our review of LayerZero’s Stargate, we identified a business logic bug that could potentially lead to desynchronisation of token balances in swaps between two coins. This error would have broken the Instant Finality Guarantee and led to user funds being permanently locked.
Outside of a formal security review, one of our security researchers also discovered a critical vulnerability ‘in the wild’ in Premia, one of Arbitrum’s leading options protocols. An allowance check issue would equip any user to grant allowance to themself to arbitrarily cause cross-chain transfers of other users’ tokens to an arbitrary address i.e. any user can steal any other user’s funds using cross-chain transfers.
For Arbitrum DAO in particular, security considerations for governance proposals are two-fold:
- Are the proposed changes secure? Zellic’s deep expertise across the full stack of blockchain systems will allow us to perform a thorough manual review of every change.
- Do the changes introduce new vulnerabilities and/or attack vectors? I.e., are there dormant backdoors in the proposal? Zellic will develop threat models for all relevant governance proposals to ensure that proposed changes are secure and do not implement soft-backdoors that can be activated at a later date. An important example here was the self-granting of 1.2M votes on Tornado Cash by a malicious actor.
Project Management
In the last two years, Zellic has worked with 200+ clients on security reviews that last as little as one day to longer than twelve months. We follow strict standard operating procedures to ensure the highest quality security services and communication with our clients. We assign at least two auditors and one engagement manager for every project. Our engagement managers act as a further layer of quality assurance for our clients, ensuring our work always meets our standards.
Of particular relevance to Arbitrum DAO, we have worked closely with several L1 and L2 teams and foundations including Solana, Sui, Aptos, TON, Osmosis, Mantle, Scroll, and StarkWare.
Purpose/Mandate of the ARDC
The purpose of the ARDC is to provide objective reporting on governance proposals and to support making the DAO’s ecosystem visions a reality. We have outlined our proposed contributions to ARDC’s mandate in the section titled Proposal ‘Review & Assistance’.
Additional Contributions
Scope of Services and Applicable Fees
We take accountability seriously, and only expect the DAO to compensate for work that has been satisfactorily completed. We will define clear milestones for each initiative, and only request payment upon completion of a milestone. We are committed to the Arbitrum ecosystem for the long-term, and request that our ARB payment be locked up according to a predetermined schedule. We leave the decision of the vesting period up to you, the delegates, but suggest a minimum period of 6 months. Overall, we expect our contributions to positively impact the ecosystem and are confident that it will be reflected in the long-term value of $ARB.
The scope of services listed below is an outline of possible deliverables to expect from Zellic. We do not presume that all will be within scope or a priority to the DAO.
Rather than a prescriptive approach, we want you, the delegates, to pick and choose what you find useful.
| Description | Required time (engineer-weeks) | Cost (USD) | Cost (ARB) | Notes |
|---|---|---|---|---|
| Template Primitives for Arbitrum Stylus | 6 | $135,000 | 65,534 | An ERC20 implementation for Stylus will take ~3 eng weeks as a benchmark. We can build 2 templates based on the DAO’s priorities. |
| Guidelines and frameworks for AIPs | 8 | $180,000 | 87,378 | Two security researchers will work in tandem for four weeks. |
| Arbitrum CTF Competitions | 10 | $225,000 | 109,223 | Time includes writing challenges, managing infrastructure, and organizing the competition. |
| Security primers on Arbitrum | 3 | $67,500 | 32,766 | We plan to write one primer every two months, based on the needs of the DAO. Each primer will take one engineer week. |
| Governance attack simulations | 6 | $135,000 | 65,534 | Two security researchers will work in tandem for three weeks. |
| Arbitrum Drift Tracker | 6 | $135,000 | 65,534 | Developed by Zellic’s engineering team. |
| Forky | 4 | $90,000 | 43,689 | Developed by Zellic’s engineering team. |
Again, we urge delegates to pick and choose only the components in this proposal that would be valuable to the goals of the ARDC.
Some qualifications on the pricing above:
- ARB price was calculated at $2.06 as of February 12, 2024.
- Our market rate for security reviews and advisory is $25,000 per engineer-week. Given our commitment to DAO security and Arbitrum-at-large, we are extending a flat 10% discount for services outlined above, at a rate of $22,500 per engineer week.
Voluntary Token Bond
If selected for the ARDC, Zellic will at our own expense post a $75,000 USD bond. This bond will be used to buy ARB tokens which will be staked for the full duration of the performance of this proposal, up to a maximum of 12 months. The ARB tokens will be returned to us after this lock-up period.
Summary
In preparing and drafting this proposal, Zellic has been grateful for the openness and transparency of various Arbitrum DAO stakeholders on the subject of core protocol developments, security considerations, and ARDC priorities. As such, we have made considerable effort to be precise with our scope of work to address the specific needs of the Arbitrum ecosystem.
Zellic’s commitment to blockchain security is deeply aligned with Arbitrum’s work as a forerunner in securely scaling Ethereum. Zellic appreciates the opportunity to submit a proposal for the Security Member in ARDC, and looks forward to a continuous partnership with the DAO and its delegates. We thank Arbitrum DAO for its consideration.
Feel free to attach any relevant documents, portfolios, or links to previous work or contributions.
You can find our publicly available audit reports here: GitHub - Zellic/publications: Zellic's audits, publications, and reports.