Reproduce “GTDD: Generative Test-Driven Development for AI Coding Agents with Adversarial Testing” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Masahiro Kato

GTDD separates development testing from acceptance audits. After a candidate is committed, a separate tester proposes inputs; a trusted evaluator checks them against a reference and returns simplified failures for repair. In a key-value-store study, Dynamic informed had a lower final mean failure rate than Static LLM, 0.274 versus 0.405, but the comparison does not isolate regeneration’s effect.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.