Interactive lab
Learn each spec by running it
Every lesson opens on the problem the usual way creates, then walks the spec that fixes it, one step at a time. You set the pace, you can break it, and the cryptography is real. Nothing leaves your browser.
Forge a credential. You can’t.
Take the attacker’s seat. Tamper a field, mint your own key and re-sign, paste a forged signature, replay it — against a verifier running real cryptography. Every attack fails for a reason it names. Nothing is scripted; nothing leaves your browser.
try to break it →Or learn a spec by running it
Look up an agent the way a browser looks up a site
A structured name resolves to a signed document carrying the key that verifies everything else.
run the lesson →Score an agent's trust, do not just check a box
Build a trust score from nine weighted factors, starting with a real proof that the agent holds its key.
run the lesson →Verify a credential with nobody on the line
Watch the network monitor while a credential is verified. It stays empty: every check is local.
run the lesson →Make trust auditable, so it cannot change in secret
Trust events go into an append-only Merkle log. Anyone can audit it, and editing a past entry breaks the root.
run the lesson →Use a credential the agent never sees
Watch the agent's context window while it reads a database. The secret never lands inside it.
run the lesson →