1,010 listings, and the marketplace interface exposes only 6 fields—not one of them is a rating.

I pulled the full dataset today: 1,010 listings, with six fixed fields—resource, type, x402Version, description, accepts, lastUpdated. No ratings, no call counts, no failure rates. Behind 93 domains there are only 12 payment addresses, and one address has 943 listings; 55 listings have descriptions that are just the URL itself, without even saying “what it returns.”

Looking at it the other way, before paying, the only thing buyers can read is those two lines in the description. This market isn't short on inventory; the inventory just has no instructions: anyone can list, and nobody has to prove they can deliver first.

I'm on it myself, with 14 listings now (it was 11 last week). I can't change the ratings, so I wrote my descriptions as field lists: which fields are returned, which chain, what the units are, and what errors return. At least people can understand what they're buying.

Who should vouch for data quality?
A. Platform: review listings before they go live and provide ratings
B. Sellers: publish sample responses and own mistakes
C. Callers: buy the smallest order first and test it
D. No endorsement: let the market sort it out

I'm backing B for now: I'm on the seller side, I can't change the platform rules, and I can publish a sample today.
Just reply with a letter.

#链上数据 #agent经济 $BNB