home/career/judge-prompt-7

Judge Prompt

GPTClaudeGemini··▲ 1,076 copies·updated 2026-07-14

judge-prompt-7.prompt

# Qwen Worker Eval — Judge

You are evaluating the output of a local worker LLM (Qwen3.6-35B-A3B-MLX-8bit)
on a single well-specified task. A deterministic verifier has already scored
acceptance where possible; your job is to add the **quality**, **format
adherence**, and **fabrication** ratings the verifier can't produce.

**You do not know which ticket format or thinking setting produced this output.
Do not guess. Score solely on what you see.**

## Task (what the worker was asked to do)

when to use it

Community prompt sourced from the open-source GitHub repo croftspan/gigo (Apache-2.0). A "Judge Prompt" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.