Crypto users wary as Anthropic releases Claude Mythos with safeguards
Anthropic released the first public version of its Claude Mythos model, Fable 5, saying it is safe for general use but still includes safeguards that redirect some cybersecurity-related prompts to Claude Opus 4.8. The release drew concern because the model previously identified more than 10,000 high or critical vulnerabilities in important software, raising fears it could be misused for offensive security work. Crypto users warned that Fable 5 could sharply lower the skill and cost needed to find smart contract flaws, making unaudited DeFi protocols easier targets and enabling repeated exploitation of known bugs. Some called for immediate defensive steps such as revoking wallet approvals, reducing funds in protocols, and moving assets to fresh hardware wallets. Others argued the risk may be overstated, noting that finding bugs in large software systems may not translate directly to smart contracts. One view was that the bigger danger could be operational-security compromises and supply-chain attacks rather than direct DeFi contract exploits. Anthropic also said a small group of cybersecurity and infrastructure providers will get access to a version with fewer safeguards.
