Benchmark reproduction
Run the three published known-answer scenarios, preserve the environment and outputs, then submit the complete public evidence package.
Reproduce the benchmark29 / Independent evaluation
PactVerity invites developers, security reviewers, researchers and technical writers to reproduce the public evidence stack and publish both successful and failed results. No token purchase, wallet connection or promotional commitment is required.
01 / Evaluation routes
Each route starts from public inputs and requires the reviewer to preserve commands, versions, outputs, limitations and any material connection to PactVerity.
Run the three published known-answer scenarios, preserve the environment and outputs, then submit the complete public evidence package.
Reproduce the benchmarkReview what a receipt proves, identify a concrete limitation or failure mode and publish a reproducible finding without exposing secrets.
Open review requestUse a public API or agent workflow, state the measurement source and publish the exact receipt, verification result and limitations.
Choose an integration trackDownload the dataset, citation files and methodology. Cite the version used and disclose modifications or first-party relationships.
Open citation kit02 / Evidence package
Name the PactVerity page, specification, source revision, API version or dataset version being tested.
Publish the runtime, command, timestamp, public input sources and expected result without credentials or private data.
Publish complete machine-readable results, receipt digests, failures and deviations from the documented behaviour.
State any payment, token holding, employment, sponsorship or other material connection. No relationship is assumed from participation.
03 / Publication rule
A technically useful negative result is more valuable than a promotional mention. PactVerity will not request favourable conclusions or suppress a reproducible limitation.
Public review queue