Audit Comp | Shardeum: Core III

Which chains and/or networks will the code in scope be deployed to?

Shardeum

Which parts of the code are you most concerned about? We are concerned with the web3 and business logic within all four repositories in this boost. Things like transaction queuing, penalties, and consensus. This includes any internal transactions or things involving the global account.

What attack vectors are you most concerned about?

Parsing/signature errors, cheating the rotation system, and transaction processing. We received quite a few message parsing and signature related reports in the previous boosts and feel like there may still be some vulns to find. Secure accounts and multisig transactions involving them will be valuable targets and need extra scrutiny. Which part(s) of the system do you want whitehats to attempt to break the most? transaction queuing, penalties, and consensus. Are there any assumed invariants that you want whitehats to attempt to break? Sum of EOA account balances before attack == Sum of EOA account balances after attack + transaction fees. This should cover SHM disappearing from the network or being created out of thin air

What external dependencies are there?

These are listed in package.json

Where might Security Researchers confuse out-of-scope code to be in-scope?

A note on Shardeum and Shardus Core scope: the default config in release mode in the branch is in scope. Whitehats are free to configure, patch, and modify their own malicious nodes however they want. However, target nodes must be running the default config in the target branch in release mode. This is to prevent the whitehats from wasting time reporting things we specifically allow in debug mode. The only exception is minNodes and maxNodes settings, which allow different size networks to be created. Certain vulnerabilities may only exist in certain network sizes, and we do not wish to limit Whitehat activity and participation for lack of computing power attempting to run a large local network. However, network-wide attacks that only work under 128 nodes may be rejected or reduced in severity at our discretion. If the researchers can enable debug mode options remotely then that is valid and can be paid out.

Attacks that require the network to still be initializing/bootstrapping are out of scope. Wait until the network mode reaches “processing” + 15 cycles after startup before launching attacks. The rules for staking/join are a little different and the network will not be public during this time. Attacks on a network that is repairing itself (was once in “processing” mode but has since degraded to “safety” or “recovery”) are in scope. This bounty introduces the concept of a KYC-required “genesis node”. Attacks performed with genesis nodes are in scope, attacks performed against genesis nodes are in scope. Nodes attacking themselves are out of scope.

0day vulnerabilities in dependencies are in scope. Any other vuln in dependencies is out of scope. Smart contracts are out of scope

Finally, the more nodes that are required to launch an attack, the more at risk the vuln is of being downgraded. If it takes 33% (for example) of the nodes in the network being malicious to cause damage, then it becomes difficult to distinguish the impact from a brute-force/51% attack, which is completely out of scope.

Are there any unusual points about your protocol that may confuse Security Researchers?

Please consider how your vulnerability will behave on a network with a shard size of 129 nodes. We will accept reports with a PoC on a smaller network, but the severity may be affected if the impact is less feasible on network with a shard size of 129 nodes.