Anthropic's Claude Mythos fuels AI script‑kiddie surge
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic released Claude Opus 4.7 with built‑in safeguards to block malicious cybersecurity requests after unveiling Claude Mythos.
- DARPA hosted the AIxCC in Las Vegas where teams scanned 54 million lines of code, finding dozens of real bugs beyond the injected ones.
- Trail of Bits CEO Dan Guido warned of a “tidal wave” of AI‑enhanced script kiddies and called 2026 the make‑or‑break year for security debt.
- XBOW topped HackerOne’s bug‑bounty leaderboard in June 2025, beating human hackers and showing AI’s offensive capability.
- Project Glasswing is part of Anthropic’s effort to prevent misuse of its AI tools.
Why it matters: Software vendors lose billions in breach costs as AI‑powered script kiddies multiply; attackers gain cheap zero‑day access, and security teams must invest heavily before the 2026 security‑debt deadline, inflating remediation budgets.
