home/roleplay/verification-rubrics

Verification Rubrics

GPTClaudeGemini··▲ 908 copies·updated 2026-07-14

verification-rubrics.prompt

# Verification rubric prompts (LLM-as-judge)

When a deterministic check (tests, lint, schema) can't capture what "good" means
— tone, UX quality, "is the framing right for this audience" — you reach for an
**LLM-as-judge**: a model that grades output against a rubric and returns a
verdict with evidence. Use it *in addition to* deterministic checks, never
instead of them, and always ground the verdict in evidence.

→ Background: [docs/05 — Stop Conditions & Verification](../docs/05-stop-conditions-and-verification.md)
(including the known biases of LLM-as-judge).

> Every prompt here is written to copy and paste. Primary sources are linked
> inline where a specific claim or technique depends on one.

---

## A binary PASS/FAIL grader (the simplest useful shape)

when to use it

Community prompt sourced from the open-source GitHub repo invincible04/awesome-loop-engineering (MIT). A "Verification Rubrics" 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.