Welcome to Tech Beat, your daily read on the stories shaping our digital world.
A significant security finding is making waves in developer circles today. Researchers at Apiiro discovered that Claude Code, Anthropic's agentic coding tool, has been quietly leaking sensitive credentials into public repositories. The culprit is a convenience feature — a local settings file used to remember user preferences — which inadvertently exposed over two thousand secrets including API keys and tokens sitting in plain sight on GitHub.
That story connects to a broader tension playing out in financial services, where major US banks are now drawing a hard line against AI agents accessing customer accounts. Institutions are not only blocking third-party bots from logging in on users' behalf, they're also making clear that if an automated agent makes a costly mistake with your money, the liability lands squarely on you — not the bank.
And in the Baltic Sea, a Russian cargo vessel came under drone attack off the coast of Finland today, adding another chapter to the ongoing pattern of maritime incidents in those waters. Details remain limited, but the strike underscores how drone warfare has extended well beyond land-based conflict into critical shipping lanes in northern Europe.
Those are the stories worth watching today. Keep surfing. Tech Beat out.
