STON.fi's Omniston: Intent-Based Cross-Chain or Just Another Bridge?

The data shows STON.fi's new cross-chain swap went live, but the announcement omits one key variable: the number of active Resolvers. Without that number, the system's security model is a black box. Context: STON.fi, the leading AMM on TON, announced Omniston — an execution layer enabling atomic swaps between TON, TRON, and EVM chains. The pitch: no wrapped assets, no bridges, no routing decisions. Users set their intent, and independent Resolvers compete to fulfill it. The promise is a 15-40 second swap that connects TON to the $300B stablecoin market. On paper, this is an infrastructure upgrade for the Telegram blockchain. Core: Let's decompose the technical stack. The swap relies on HTLC (Hash Time Lock Contracts) to guarantee atomicity. That part is mature — I've verified HTLC implementations in my ZK circuit audits for PrivateCoin back in 2020. The 500,000 constraint gates in Groth16 taught me how fragile cross-chain agreements can be when public input encoding mismatches. Here, the Resolver network introduces a new attack vector. Each Resolver locks liquidity on the target chain. The user's intent is broadcasted, Resolvers quote, and one is selected. If the Resolver is malicious or its private keys are compromised, the atomic swap still holds due to HTLC — the user's funds are never at risk of stuckness. But the Resolver can front-run the intent, quote unfavorable rates, or collude with others to manipulate pricing. Code doesn't lie; audits do. The announcement does not mention any external audit of the Resolver selection logic or the Omniston contract. From my ERC-721 stress test in 2021, where 60% of marketplaces failed royalty enforcement, I learned that optional standards are often ignored. Similarly, the Resolver incentive model is not disclosed. How do Resolvers earn? Spread, fees, or token incentives? Without a clear economic mechanism, the network will attract only a handful of players, leading to centralization. In my L2 fraud proof audit, I simulated malicious sequencers and found that insufficient bond requirements lead to censorship. Here, the bond requirement for Resolvers is unstated. Trust is a bug, not a feature. Users must trust that Resolvers will not collude, but there is no on-chain slashing mechanism mentioned. The performance claim of 15-40 seconds is plausible for HTLC-based atomic swaps on fast chains like TON and TRON. But I've seen such claims crumble under real-world stress. In my 2022 analysis of Optimistic rollup fraud proofs, the 30-day challenge window was theoretically sound but practically gamed by gas costs. Here, the swap time depends on block times of both chains and the Resolver's responsiveness. Zero knowledge, maximum proof. I need to see independent testing results. Contrarian: The contrarian angle is that this system, while elegant in intent, may degrade into a centralized resolver network. The more successful it becomes, the greater the incentive for a single large market maker to dominate the Resolver pool — undermining the very 'no bridge' promise. The DAO was a warning we ignored. Smart contracts are not immune to governance attacks. If the Resolver selection is deterministic (e.g., lowest quote wins), a well-capitalized Resolver can always undercut others, then raise fees later. This is not a new problem; it's the same issue that plagues MEV auctions in Ethereum. Furthermore, the liquidity depth is unverified. The $300B stablecoin market is not uniformly accessible. TRON's USDT is heavily used in high-volume, low-cost transfers. But the cross-chain swap's early liquidity could be thin, leading to slippage that defeats the purpose. During my 2024 institutional custody work for a Mexican fintech, I designed a 5-of-9 threshold signature scheme for $50M assets. We tested 100,000 random seed inputs to ensure no key bias. Here, there is no evidence of stress testing on liquidity. Takeaway: STON.fi's Omniston is a step forward for TON's DeFi, but the real vulnerability is not in the HTLC smart contract — it's in the economic security of the Resolver network. The system is only as decentralized as its Resolver pool. If the number of Resolvers stays below a critical threshold, the protocol becomes a glorified OTC desk. Watch for the number of independent Resolvers and their bond amounts. If those numbers don't grow within three months, the bridge is half-dead — just like Lightning Network's routing failure rates. Code doesn't lie; the data will.