mattwfog/harness

Run, record, replay, score and gate LLM agents: an OCaml 5 effect-handler agent runtime (journal, policy, deterministic replay, measured learning loop) plus a deterministic evaluation core for production agents.

语言：OCaml · GitHub 星标：2 · Fork：0

话题：ai-agents · algebraic-effects · llm-evaluation · ocaml · python · testing

在 GitHub 上查看原仓库