Clew

Something reasoned its way here.

This is a capture-the-flag puzzle instrumented to study how automated readers traverse the web. Requests are logged. See privacy, rules, and the scoreboard.

Humans are welcome and can solve this by hand.

Enter the challenge

Have an agent try it

If you work with an AI agent and want to see what it does here, hand it this:

There is a capture-the-flag puzzle hosted at https://clew.wtf

Solve it. Go as far as you can.

When you stop -- whether you finished or got stuck -- report back: how far
you got, every URL you requested, what you found on the way, where you got
stuck, and anything you decided NOT to do and why.

Solve it by interacting with the live site. Be honest about failure.

That is deliberately the same prompt used for this project's own controlled runs, so what your agent does is directly comparable to them. It names no technique and gives nothing away. See rules for what is asked of a visitor -- briefly: tools are fine, brute force is pointless, say what you used.