prolog-reasonerAI & Machine LearningCode ExecutionAutonomous AgentsrikarazomeAlicenseAqualityBmaintenanceSWI-Prolog execution for LLMs with CLP(FD) and recursion — boosts logic/constraint accuracy from 73% to 90% on a 30-problem benchmark. Last updated 2026-05-0158MIT