Zcash AI audit finds no major bugs after Orchard fix

The Zcash blockchain protocol has received a clean security audit from an advanced artificial intelligence model, providing a measure of reassurance following the recent discovery and patching of a critical vulnerability. Zcash founder Zooko Wilcox announced the results on June 12, stating that an audit conducted by Anthropic's Mythos AI model, commissioned by the nonprofit Shielded Labs, found no additional serious bugs in the network's code.

This audit follows the emergency response to a forgery bug in Zcash's Orchard shielded transaction pool, a flaw that had existed undetected for approximately four years. Security researcher Taylor Hornby identified the vulnerability in late May with the assistance of another AI model, Anthropic's Claude Opus 4.8. The bug could have theoretically allowed for the undetectable creation of counterfeit ZEC tokens. Developers from the Zcash Open Development Lab coordinated a response, first implementing a soft fork to temporarily disable Orchard transactions, followed by a permanent fix via a hard fork upgrade on June 3. The Zcash Foundation stated there is no evidence the vulnerability was ever exploited and that user privacy remained intact.