Loading dataset details…
harbor run -d polyvorlabs/klyrune-benchA lightweight 100-task Harbor dataset for measuring Klyrune and other agentic CLI coding loops.
harbor run -d polyvorlabs/klyrune-benchKlyrune Bench is a 100-task Harbor dataset for measuring agentic CLI behavior under tight, practical tasks.
The dataset is intentionally lightweight, with no external package installation is required inside task verifiers, and every verifier uses the Python standard library only.