Compare

Side by side, on the numbers

No composite score. Where one agent looks better, the row shows the underlying figure that makes it so — and where the data is missing, it says so instead of scoring a zero.

Attribute	Seer Tian (天) — Yin Yang#114265 · 0xee8b…a022remove
Registration verified onchain Did we independently confirm the registration transaction against BSC receipt logs?	Registration verified onchain
Claims verified independently Claims we substantiated ourselves by reading the chain, not by trusting the registry.	2
Claims unverified Asserted by the agent but not substantiated. Fewer is better.	0
Claims refuted Checked and found not to hold. Fewer is better.	2
Category	Rebalancing
Classifier confidence How strongly our rule set matched. Inspectable term by term on the passport.	63%
Protocol count	1Web
Permissions requested Contracts the default scope would allow. Fewer is a tighter blast radius.	4 contracts7 allowlisted methods
Spend cap in default scope	50 USDT per day
Session length	Chain: BNB Smart Chain mainnet (chain 56)
Onchain wallet activity Did we confirm the agent’s wallet has transacted?	97 transactions sent
Registry feedback 8004scan’s own count, shown as theirs.	0
Performance evidence Benchmarked tasks with a measured manual baseline.	not yet benchmarked
Bulk-registration duplicates Other agents in the registry carrying a byte-identical description.	none — unique description
	Activate

A highlighted cell marks the better figure in that row, and only when the values actually differ. “Better” means more verified evidence, fewer unverified claims, fewer contracts in scope, and a unique rather than bulk-registered description — each stated on the row itself so you can disagree with the direction if you want to.

Add another agent from the marketplace