Judge the code block information in the markdown

CLI for judge-java — init, dashboard, and management commands

Agent evaluation harness. Judge sessions on success, tool usage, efficiency, methodology. Inspired by opencc.

Judge the condition

Closed-loop, evidence-gated build system for AI coding agents. Plan, Spec, Build, Verify — with a separate Judge role that holds the exit gate and issues rework orders until the frozen Definition of Done is met.

javascript type judge

Runflow Evals — project-local evals framework for Runflow agents (datasets, scorers, journey/conversation validation, LLM judge, viewer)

LLM-judged CLI test runner — define test cases in YAML, execute commands, let an LLM judge PASS/FAIL. Outputs JSONL trace.

MCP server that lets an AI agent judge whether a flight price is actually a deal, grounded in observed fare history. Answers 'is $X for this route a good price?' with real numbers an LLM cannot produce on its own.

Atbash safety judge exposed as a standalone MCP server

Eval surfaces: versioned LLM-as-judge rubric, eval-grader loop, skill golden-set runner.

Prompt templates for LLM Judge Toolkit evaluation criteria

MCP server for OpenGATE — let an AI agent check whether its answers are grounded in the provided context. Deterministic, no LLM judge.

Local MCP server implementing OpenRouter's Fusion panel architecture: fan-out + two-step judge.

Dynamic routing policies for @swoosh-dev/router — classify the prompt with an LLM judge (structured output) and route accordingly.

Optional LLM judge (WebJudge-style) for taskproof: a versioned rubric prompt that grades goal completion from a run's evidence, gated behind the deterministic assertions

A tool for Baekjoon Online Judge

Dataset regression evals for the AI SDK: pluggable scorers (exact-match, embedding similarity, LLM-as-judge) that compose with your own vitest suite

judge-solve0

Calibration metrics, runner, datasets, and drift detection for LLM Judge Toolkit