Cute Crypto 🐷

How can AI agents trade and execute safely in financial markets? Our CEO Raul Heraud explored this question with fellow panelists at @wallet’s Beyond X in Singapore. Proud to support an evening bringing this conversation to the stage.

On 5 October, the Ethereum Foundation's Trillion Dollar Security initiative published a post on native transaction assertions, with EIP-7906 as one possible design. For the Hegotá upgrade, EIP-7906 has advanced to "Considered for Inclusion" but is not yet confirmed. The idea: a transaction carries a rule about its own outcome, signed along with its actions. After the actions run, read-only code compares starting and final values against that rule: a minimum amount received, no new approvals, an account's control logic left unchanged. If the rule fails, the actions revert. That puts the check after the actions have run, where the full set of net state changes can be inspected. The post is direct about two conditions it still depends on. The first is where the rule comes from. "A compromised frontend can write an assertion that permits the attack." A useful rule, in the post's words, "must come from independently approved user intent, a standing account policy or protocol logic that the compromised component cannot change." The second is whether the rule is there at all. "EIP-7906 does not require transactions to include an assertion, so whoever builds the transaction can leave it out." A wallet has to include it in every frame transaction it builds (the transaction type EIP-7906 builds on, defined in EIP-8141); a protected function has to reject ordinary transactions and frame transactions that arrive without the exact assertion it expects. What the check protects depends on two things outside the check itself: who wrote the rule, and whether anyone can leave it out. The same two questions apply to any verification layer, including ours: whose rules it checks against, and whether it can be left out. blog.ethereum.org/en/2026/10… #DeepSafe #Web3Security