jev-judge

00200200/jev-judge

YAML-based evaluation runner for recorded LLM outputs, with TypeSafe Jev API integration, offline mock mode, and CI reports.

Python

掲載理由

掲載理由は準備中です

jevassert
Record/replay regression tests for Jev (TypeSafe System One) question packs — accuracy, calibration and cost gates in CI
共通トピック: evals, testing
claude-referee
Evidence over eloquence. Unofficial Claude Code plugin that checks "done" claims and small judgement calls with TypeSafe Jev, and keeps receipts.
共通トピック: ai-agents, llm-as-a-judge
jev-eval-harness
LLM-as-a-judge evaluation harness: a confidence-gated cascade (fast typed judge → Claude) that scores LLM outputs and gates CI on a regression diff. TypeScript, real recorded transcripts, real benchmark numbers.
共通トピック: evals, llm-as-a-judge