Invest1 publisherNot yet confirmed elsewhere2 min readPublished
Ethereum Foundation's answer to Bybit-style theft waits until 2027 at the earliest
Ethereum Foundation says EIP-7906, a check that could have reverted the $1.5 billion Bybit theft, will not ship before 2027. Its own post concedes the check cannot stop the stolen-key and social-engineering attacks behind most of 2026's losses.
The Investor · Invest desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- EIP-7906 adds a read-only check at the end of a transaction, comparing balances, storage and events against a rule and reverting the whole transaction if the rule fails.
- In March 2026 an Aave user swapped $50.4 million of aEthUSDT for roughly $36,000 of aEthAAVE, a trade a minimum-output assertion would have blocked.
- EIP-8141, the frame-transaction base the check sits on, is scheduled for the Hegota upgrade, but EIP-7906 itself has only reached Considered for Inclusion status.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- exposure Exchanges and custodians holding hot-wallet or multisig keys keep the full loss from key theft through at least 2027, because an attacker holding the keys can sign the assertion too.
- constraint Once live, the check protects only accounts that set standing rules or independently approved intent in advance, so a wallet with no policy in place gains nothing from the upgrade.
- decision The Hegota meta proposal still has to confirm EIP-7906, and the Foundation's post is the case for doing so; Cryptopolitan says it raises the odds.
CertiK counted $1.32 billion lost to crypto security incidents in the first half of 2026 [10]. TRM Labs attributes around two-thirds of that, about $880 million, to North Korea-linked groups [11][20]. In TRM's breakdown, infrastructure and operational compromises took close to five times their share of incidents in dollars [17]. Drift and Bitget fall in that category. Drift lost $285 million after a six-month social-engineering operation compromised contributors' machines, and assertions do little for a case like that [13]. Bitget lost $387.5 million in September through compromised hot-wallet keys [14]. Together the two come to $672.5 million [19].
The proposal addresses the gap between what gets signed and what actually happens. A signature authorizes a request, but the result depends on the code and state that request meets when it executes [6]. At Bybit, a masked signing interface led signers to authorize a swap of the Safe's implementation contract [7]. The Foundation says a standing rule forbidding that contract from changing would have reverted it [7]. Instead the Lazarus Group took around 400,000 ETH [8]. The Aave user kept about 0.07% of the position's value [18]. A stale gas ceiling had rejected better-priced quotes, and the winning solver failed to execute [9].
The Foundation's Trillion Dollar Security initiative wrote on October 5 about how "native transaction assertions could protect users where today's defenses stop" [4]. The Foundation also concedes the limit. A rule helps only if it comes from independently approved intent or a standing policy the attacker cannot rewrite, and an adversary who holds the keys signs the assertion too [15].
Both the date and the reach can move. If the check misses Hegota, it lands later than the earliest date the Foundation gave [2]. If attackers go back to Bybit-style interface masking, the share of stolen dollars an assertion can catch grows [7]. If they keep going after keys and machines, that share stays small, because TRM puts most of the money there [17].
The protocol work here goes into checking outcomes, and the check defers to whoever signs [15]. In my view, keys and signer machines remain the operator's problem through at least 2027 [2]. I would drop that view if TRM's next breakdown showed infrastructure and operational compromises well below the 76% of dollars they took this time [12].
What to watch
- Whether EIP-7906 moves from Considered for Inclusion to confirmed in the Hegota meta proposal.
- TRM Labs' next loss breakdown, and whether infrastructure and operational compromises keep close to 76% of dollars lost.
- Whether Safe and other multisig wallets prepare standing rules, such as locking implementation contracts, for when assertions ship.