agent-eval-harness
Python
★ 0
updated 17d ago
Evals, guardrails, and tracing around an LLM document-extraction agent. Strict tool-use boundary, provenance/totals/schema guardrails, a scored regression-gate eval over a golden set, and a structured trace per run. Runs offline in ~5s.
No plain-English explanation yet — one is being written right now. Check back in a minute.