llm-as-judge

llm-as-judge is a skill for Claude Code from shimo4228/claude-harness. It costs 116 tokens per session (2,322 once invoked), scanned A, original, MIT.

A method for designing an evaluator in which a language model answers concrete yes-or-no checks and then gives one overall named decision.

In plain words
What is it for?
Use it to design or review language-model quality gates, evaluation prompts, and verdict formats.
Why use it?
It avoids unreliable numeric scores and prevents several minor passes from hiding one serious failure.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter. Also seen: mentions subagents; mentions Claude Code.

Good fit Use it to design or review language-model quality gates, evaluation prompts, and verdict formats.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/shimo4228/claude-harness/llm-as-judge
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add shimo4228/claude-harness --skill llm-as-judge
Clone the repo
git clone --depth 1 https://github.com/shimo4228/claude-harness

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for llm-as-judge

README.md
[![agentmods](https://agentmods.dev/badge/skills/shimo4228/claude-harness/llm-as-judge/github.svg)](https://agentmods.dev/skills/shimo4228/claude-harness/llm-as-judge)
Your own site
<a href="https://agentmods.dev/skills/shimo4228/claude-harness/llm-as-judge"><img src="https://agentmods.dev/badge/skills/shimo4228/claude-harness/llm-as-judge/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for llm-as-judge

Your own site · 80×15
<a href="https://agentmods.dev/skills/shimo4228/claude-harness/llm-as-judge"><img src="https://agentmods.dev/badge/skills/shimo4228/claude-harness/llm-as-judge.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 116 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,322 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00116 $0.02322
Opus 5 $0.00058 $0.01161
Sonnet 5 $0.00023 $0.00464
Haiku 4.5 $0.00012 $0.00232

Measured 12d ago against content hash a5c8edd90c02, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

llm-as-judge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/llm-as-judge/SKILL.md · 197 lines

How it starts

The opening of the file, as written. The whole thing — 197 lines — stays where its author put it; the contents beside it link to each section on GitHub.

LLM-as-Judge — Checks as Evidence, Holistic Verdict, No Scores

Core rule in one line:

Collect evidence with binary Yes/No checks, let the LLM issue one named holistic verdict, and never aggregate the answers into a score.

Why not rubric scores

  • Irreproducible. Numeric scores on the same input drift between runs (3 vs 4 out of 5). LLMs are bad at independent per-dimension scoring — they get pulled by the overall impression, and central-tendency bias compresses scores into a narrow band that straddles any threshold you pick.
  • Dilution. A sum or satisfaction ratio converts one fatal defect ("the referenced file does not exist") into a small deduction. One dominant No must be able to decide the outcome alone.
  • Unexplainable. Nobody can say why something is a 3.5. A No answer to a concrete question explains itself and doubles as the improvement item.

Rubrics were invented to structure human evaluation, where assessors can score dimensions independently. For an LLM, invert the design: force the checks it tends to skip, and leave the judgment holistic — that is what it is good at.

The three principles

Principle Do Don't
① Binary checks Decompose criteria into Yes/No questions with 1-line evidence each "Rate specificity 1–5"
② Named holistic verdict Pick exactly one verdict from a fixed enum, judging the whole "Total 12 points → pass"
③ No aggregation Enumerate the No answers as the verdict's rationale Use the Yes-ratio as a quality metric

① Binary checks

Ask "is there a runnable command example? Yes/No", not "how specific is this?". Good binary questions are verifiable (the text settles them black-or-white) and evidentiary (a No names what is missing). Require one line of evidence per answer — a file read, a path check, a measured command output.

② Named holistic verdict

Holistic judgment — one conclusion from the whole, without passing through per-dimension numbers — is the LLM's strength. Design the verdict enum so each value maps 1:1 to a next action, e.g.:

Read the full file on GitHub · 197 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 12d ago First seen · 197 lines · 116 tokens per session scan A a5c8edd90c02

Subscribe to this mod's changes

llm-as-judge is a skill published in the GitHub repository shimo4228/claude-harness (3 stars, last pushed 6d ago), licensed MIT. It adds 116 tokens to every session and 2,322 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

ai-evaluation

Systematic evaluation (evals) for LLM and AI products. Design test cases, measure accuracy/quality, track regressions, benchmark models, and build continuous evaluation pipelines. Distinct from traditional software testing with probabilistic outputs.

sunnypatneedi/claude-starter-kit · 51 tokens

evaluating-llms-harness

Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.

NousResearch/hermes-agent · 29 tokens

cli-eval

Create and run evaluation suites, watch live benchmark progress, view scorecards, compare model performance, and integrate eval runs with CI workflows from the CLI.

diegosouzapw/OmniRoute · 34 tokens

model-merging

Merge multiple fine-tuned models using mergekit to combine capabilities without retraining. Use when creating specialized models by blending domain-specific expertise (math + coding + chat), improving performance beyond single models, or experimenting rapidly with model variants. Covers SLERP, TIES-Merging, DARE, Task…

davila7/claude-code-templates · 73 tokens

data-quality-frameworks

Implement data quality validation with Great Expectations, dbt tests, and data contracts. Use when building data quality pipelines, implementing validation rules, or establishing data contracts.

wshobson/agents · 37 tokens

dbt-transformation-patterns

Master dbt (data build tool) for analytics engineering with model organization, testing, documentation, and incremental strategies. Use when building data transformations, creating data models, or implementing analytics engineering best practices.

wshobson/agents · 47 tokens