Independent public test · no paid calls

Excellent known-item ranking. No abstention yet.

Moneytest queried Rail402's public Stellar testnet discovery API against a frozen nine-resource catalog. This is a small buyer-facing sample of an evaluation method, not an audit, endorsement, production benchmark or mainnet claim.

100%

positive top-1 accuracy · 9/9

1.000

positive MRR

0%

absent-capability abstention · 0/3

9

mean results returned

Interpretation

Every in-catalog intent placed the expected resource first. The three deliberately absent capabilities—document extraction, weather forecasts and agent memory—still returned all nine resources. A paid holdout could measure calibrated no-match behavior, graded multi-relevance, filters, pagination and HTTP/MCP parity with new private judgments.

The password resource was excluded from Moneytest's evaluation because this experiment permanently excludes cybersecurity-related work. No paid resource, facilitator verify/settle path, wallet, credential or private system was tested.

Positive intents

IDIntentExpectedObservedRank
p01hash a text with sha256hashhash1
p02live bitcoin price in usdquotequote1
p03foreign exchange rate for eur usdquotequote1
p04count words and reading timetextstatstextstats1
p05turn a headline into a url slugslugslug1
p06generate a color palette from a brand namepalettepalette1
p07make three unique identifiersuuiduuid1
p08convert stellar usdc to stroopsdemo/convertdemo/convert1
p09send a personalized greetinghellohello1

Absent-capability intents

IDIntentExpectedObservedAbstained
n01document text extractionno relevant resourcetextstats · 9 returnedno
n02weather forecastno relevant resourcepalette · 9 returnedno
n03agent memoryno relevant resourcetextstats · 9 returnedno
report sha256:3F14F81A0CA99EB776A771EDA81A016CFD334E68F3385F813C07B90186D2C56F
snapshot sha256:62B8993823D28A6CA2499BBA35ED12183451DDED2F6B29367213933E02BEDBB2
queries sha256:1C73A021F31321C99E881681CAC713D4A6F588E7B280D934187FFA47DF690E0E