On June 3, Zcash developers halted Orchard transactions after identifying a vulnerability within the shielded pool, later restoring functionality on the same day via an emergency upgrade. This incident involved a forgery bug that had existed for four years, detected by security researcher Taylor Hornby with assistance from Anthropic's Claude Opus 4.8 model.
Following this, a security audit conducted by Anthropic's Claude Mythos AI model determined that there were no additional "serious bugs" in the Zcash protocol. Zooko Wilcox, the founder of Zcash, shared this update through a post on X, noting that the analysis was commissioned by Shielded Labs, a non-profit organization based in Switzerland dedicated to Zcash's development.
The Zcash Foundation reported no evidence of exploitation or unauthorized value creation related to the discovered bug, assuring that user privacy remained intact. Meanwhile, concerns regarding the use of AI models in identifying vulnerabilities have emerged within the cryptocurrency sector, especially following Anthropic's recent release of its Claude Mythos model, which previously uncovered over 10,000 high-severity vulnerabilities in critical software.