🚨 ledger library confirmed compromised and replaced with a drainer. wait out interacting with any dapps till things become clearer. cdn.jsdelivr.net/npm/@ledgerhq/…
seems ledger connect-kit-loader is also vulnerable since it specified the dep loosely
ledger asks to use connect-kit loader to load connect-kit, but even if you follow the best practices and pin the version of the loader loader, it fetches the latest version of connect-kit >=1.0.0, <2.0.0.
this has allowed the attackers to infiltrate a shitton of libraries by compromising just the connect-kit. last known version coming from ledger is 1.1.4. three releases up to 1.1.7 were posted today, all should be considered compromised.
ledger pushed a new version, @1 should match 1.1.8 now.
it still doesn't add strict version pinning, which is better addressed here.
after investigating more than 90 million deposit and withdrawal events, i've found a supply discrepancy between the total supply weth contract reports and the actual outstanding weth.
it appears the contract holds 1 wei more than it owes. how is it possible?
the contract mints a token when you send ether to it. it doesn't track the token supply, using its own ether balance instead.
but there are two ways to send ether without triggering a contract.
first, you can set it as a block reward recipient. it will update the state, but won't run the contract code.
second, what our hacker has used, is selfdestruct (soon to be sendall). it destroys the contract and sends all ether balance to another address, without alerting it.
here is an idea for the fed devs, why not also ban 401,548 addresses which use exactly the same unverified (must be hiding something) bytecode? surely this code is up to no good. might be the same guy even.
It appears oasis.app, following Uniswap, has started sending all your data to TRM Labs. This is what happens when you connect with an address they don't like. No way to close positions from the UI, no explanation or anything.