The vibes are off for open-source devs as hackers get smarter, but the plot thickens with a new play from Anthropic. The AI lab just launched OSS Scanner, a program that lets projects tap into their most powerful models—like Claude Mythos—to hunt for security bugs in their code.
Real talk: crypto firms are highkey feeling the heat. With bad actors reportedly using AI to cook up exploits faster than teams can patch them, developers are scrambling to level the playing field. Companies like Bitcoin swap provider Boltz and payment service PayPerQ have already faced AI-assisted attacks this year.
Who is signing up?
Crypto-native teams are moving fast to secure their bags. Ethereum client developer Nethermind has already applied for audits on its entire repository, while ZEUS, a self-custodial Bitcoin and Lightning wallet, is looking to lock down its payment flows and private keys. Another notable applicant is VirtEngine, a decentralized cloud marketplace built on Cosmos.
Anthropic says the manual review process for bugs is usually a total drag, which is why they are automating the delivery of these reports. Projects will be vetted case-by-case, with priority going to infrastructure that holds the most weight in the ecosystem.
Why it matters
AI is currently a weapon that favors the attacker because finding flaws is getting cheaper while fixing them remains a slow, human-heavy process. This tool is a massive W for developers looking to build on-chain defenses that actually stand a chance. Just remember: using a new security tool isn't a silver bullet, and this piece is definitely not financial advice. Stay safe out there.
