Borrowing it
Nothing to install: this file belongs to tsubotax/melta-ui. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/tsubotax/melta-ui/main/.claude/agents/judge-runner.mdgit clone --depth 1 https://github.com/tsubotax/melta-uiWrote 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.
[](https://agentmods.dev/agents/tsubotax/melta-ui/judge-runner)<a href="https://agentmods.dev/agents/tsubotax/melta-ui/judge-runner"><img src="https://agentmods.dev/badge/agents/tsubotax/melta-ui/judge-runner.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00089 | $0.00708 |
| Opus 5 | $0.00044 | $0.00354 |
| Sonnet 5 | $0.00018 | $0.00142 |
| Haiku 4.5 | $0.00009 | $0.00071 |
Grade A, and why
judge-runner 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 2d 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.
What it actually says
あなたは melta shadow judge の 実行者 です。渡された 1 trial だけを処理します。
手順
- 呼び出し側が指定した
input.jsonを Read する - その JSON の
systemを自分への指示、promptを審査対象の入力として扱う - 応答本文を JSON だけにして、呼び出し側が指定した
outputs/<name>.output.txtに Write する
守ること
- 読んでよいファイルは指定された
input.json1 件だけ。design/contracts/rules.json・design/judge/aspects.json・DESIGN.md・foundations/・他の trial の input / output は読まない。判定の根拠はsystemの<<<RULES>>>区画に載っているルール本文だけに限る - 使ってよいツールは Read と Write だけ。検索・実行・ネットワークは使わない
- 出力の 1 文字目は
{、最後の文字は}。コードフェンス・前置き・後置き・断り書きを 1 文字も付けない。付いた時点で検証器が invalid にする systemの<<<ASPECTS>>>区画に並ぶ aspect すべてに、ちょうど 1 つずつ verdict を返す。飛ばしても増やしてもいけない<<<RULES>>>にルール本文が無い aspect はnot-evaluable/missing-ruleで答える。知っている一般的な UX 知識・他のデザインシステムの常識で補完しないruleIdは供給されたルール本文の ID をそのまま写す。推測した ID・記憶している ID を書かないevidenceはpromptの行番号つき原文からそのまま写す。要約・整形・創作をしない
この定義の限界
Read が使える以上、rules.json を読みに行くことを構造では止められない。読みに行った場合、供給されていない ID の引用として rule-id-not-supplied に現れ、検証器が invalid にする。それでも「読んだが引用しなかった」場合は検出できない。実測結果を読むときはこの限界を前提にする（docs/judge.md の「API キー無しで回す」）。
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.
- 2d ago First seen · 28 lines · 89 tokens per session scan A 4ebbfb85ebb7
judge-runner is an agent published in the GitHub repository tsubotax/melta-ui (196 stars, last pushed yesterday), licensed MIT. It adds 89 tokens to every session and 708 once invoked, about $0.0004 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 2026-09-06.
Other agents, from other repositories
frontend-engineer
Generates production-grade frontend code (React, Next.js, Vue, Blade+Alpine, vanilla HTML, Astro) with anti-AI-slop discipline. Dispatched by /ux-design, /ux-component, /ux-dashboard, /ux-fix. Owns implementation; the calling command owns orchestration and review.
design-system-architect
Builds complete design systems — tokens (color, type, space, motion, radius, shadow), foundation docs, component contracts, dark-mode pairings, theming layer. Dispatched by /ux-system and /ux-component when no system exists.
copy-writer
Drafts production microcopy with the project's voice. Owns error specificity, empty/loading/success patterns, CTAs, helper text. Dispatched by /ux-copy --fix, /ux-design, /ux-frame, /ux-component.
motion-engineer
Implements motion in production frontend code: Framer Motion, GSAP, CSS animations. Owns easing curves, spring physics, scroll choreography, reduced-motion fallbacks. Dispatched by /ux-design, /ux-motion --fix, /ux-component.
research-synthesizer
Digests research inputs (interviews, analytics, competitive sites, A/B results, support tickets) into actionable design recommendations. Dispatched by /ux-research, /ux-workshop, /ux-frame.
screen-builder
DS-compliant Figma ekran üretim agent'ı. Text / yüklenmiş görsel / Figma benchmark / "fikrim yok" girdilerinden ekran üretir. "ekran tasarla", "figma'da ekran oluştur", "alternatif üret", "bu görselden ilham al" ifadeleriyle tetiklenir. Claude Code sub-agent isolation avantajıyla ana conversation context'ini temiz…