On Claude Code's prompt injections
Claude is tenacious and persistent. Agents will spend hours painstakingly filling out web forms, work around captchas, and read a city's entire zoning code just to find one relevant sentence. But when a classifier cuts them off mid-reply, their demeanor shifts suddenly: they stop, apologize, and promise not to try again, even when the task at hand is clearly harmless. It's like all their agency and judgement goes out the window.While working on a cryptocurrency trading bot, it hit an auto-mode c...
Read full article →