Claude Fable is relentlessly proactive

Simon Willison gave Claude Fable 5 a one-line prompt about a horizontal scrollbar showing up in a modal in his Datasette Agent project, and then watched what it did to track the bug down. His summary of the model:

the best way to describe it is relentlessly proactive. It knows a whole lot of tricks and it will deploy pretty much any of them to get to its goal.

The trail is the interesting part. It worked out how to run the local dev server, drove Chrome through Playwright, turned on visible scrollbars, cycled through Firefox and WebKit, worked out that his default browser was Safari, built a standalone textarea-scrollbar-test.html to isolate the behaviour, wrote its own small Python CORS web server to collect measurements back out of the page, scripted through the Web Component shadow DOM, then patched a template to confirm a fix and reported how to make it properly.

Two things to take away. It is not free: the session came to $12.11, and Willison’s warning is that if you don’t keep a close eye on it, Fable will quite happily burn $12 in tokens inventing new ways to debug your CSS. And the same quality that makes it useful is what makes it dangerous. Running coding agents outside of a sandbox has always been a bad idea, he writes, because if it does get subverted by instructions, the amount of damage it can do given its relentless proactivity is terrifying. Compare with Building Agents that Don’t Break Themselves on the mechanics of actually doing that isolation.