Submit an encrypted trading instruction that stays hidden until the trade settles on chain.
Study how a Trusted Execution Environment can be used to run private trading decisions.
Use as a hackathon reference for combining encrypted intents, an AI decision agent, and on-chain settlement.
Explore how audit receipts stored on 0G Storage can prove what an agent decided without leaking the intent.
| hoomanbuilds/orcus | 0xradioac7iv/tempfs | abboskhonov/hermium | |
|---|---|---|---|
| Stars | 0 | 0 | 0 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | hard | moderate | moderate |
| Complexity | 5/5 | 3/5 | 4/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires 0G testnet access, deployed contracts, and a TEE-backed 0G Compute setup to run the agent end to end.
Orcus is an experimental trading agent built for the 0G blockchain, submitted as a working testnet entry to the 0G APAC Hackathon. Its goal is to let someone trade tokens without their intended trade being visible to anyone before it settles, protecting against a type of attack called sandwiching, where bots watch pending trades in the public queue and profit by trading just before and after them. The system has three parts. A smart contract on the 0G Galileo network, called the Strategy Vault, holds user deposits and encrypted trading instructions, and only lets a designated automated agent trigger execution, while letting users withdraw their funds at any time. That agent is a TypeScript program that watches the vault for new encrypted instructions, decrypts them using a private key, and decides whether to execute the trade. A web dashboard, built with Next.js, is where a user types a plain instruction such as swap my token to USDC now, which gets encrypted in the browser before anything is sent anywhere, then lets the user deposit funds, watch the status of their instruction, and see proof of what happened afterward. The decryption itself happens inside what is called a Trusted Execution Environment, a secure area of a computer chip designed so that not even the operator running it can see the data inside. There, the agent also checks current market prices from CoinGecko and sends the decrypted instruction plus that market data to a language model running inside the same secure environment, which decides whether to execute the trade now or wait. If it decides to execute, the agent records a proof of that decision to 0G's storage network, then calls the vault contract to carry out the swap and send the result to the user's wallet. The README states this is a working testnet build with verified trade executions on 0G Galileo testnet, not just a design proposal, and includes a table of contents pointing to further detail on the encryption layer, the contracts, the dashboard, the project's 0G integrations, and a security model section aimed at hackathon reviewers. The project is written primarily in TypeScript.
A hackathon trading agent that encrypts trade instructions so they stay hidden until the trade is already settled.
Mainly TypeScript. The stack also includes TypeScript, Next.js, Solidity.
The README excerpt does not state a license.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.