Evidence from OC Labs
Work you can inspect.
Published records help you assess what was tested, what the results support, and where the limits remain.
Accepted open-source contributionA database mismatch.
A check that catches it.
Micopay accepted our schema-parity regression check and decimal-column correction. Read what changed, the supporting checks and the limits of the result.
Reviewed. Merged. Public.The code and maintainer discussion are available to inspect. This is engineering evidence from an open-source contribution, not a paid-client testimonial.
Ask about your issue → Controlled software benchmark
Seven published verification cases
The published GDN evidence set records 6 verified repairs out of 7 cases, compared with 3 out of 7 for matched single-agent baselines. It includes three GDN wins, four ties and no GDN losses. One tie was a failure for both approaches.
Scope: seven small Node.js and TypeScript utility-library defects. The coordinator prepared the controls and classified the results; there was no external human adjudicator. These results do not establish performance across all software, hardware or production systems.
Open the public evidence repository →
Follow the record
The linked records contain the supporting method and limitations. Open-source project names and contributions do not imply a customer relationship or endorsement.