Dokimos

github.com ↗

Evaluation framework for LLM and AI-agent applications that scores responses, validates tool calls and execution traces, and catches quality regressions in CI.