webtoon-harness: Agent for Claude Code

.claude/agents/panel-validator.md

panel-validator is an agent for Claude Code from revfactory/webtoon-harness. It costs 161 tokens per session (3,513 once invoked), scanned A, original, MIT.

A webtoon panel checker that reviews each rendered panel for visual consistency, readable and accurate Korean text, dialogue flow, and technical problems.

In plain words
What is it for?
It is for reviewing character references, locations, speech bubbles, Korean lettering, prompt adherence, and image integrity one panel at a time.
Why use it?
Image generation can change faces, settings, lettering, or other details even with the same prompt. Checking each panel immediately identifies panels that need regeneration before they are assembled.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; mentions Codex.

This is revfactory/webtoon-harness's own configuration. It tells Claude Code how to work on webtoon-harness itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything webtoon-harness configures →

Reuse

Borrowing it

Nothing to install: this file belongs to revfactory/webtoon-harness. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/revfactory/webtoon-harness/main/.claude/agents/panel-validator.md
Clone the repo
git clone --depth 1 https://github.com/revfactory/webtoon-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 panel-validator

README.md
[![agentmods](https://agentmods.dev/badge/agents/revfactory/webtoon-harness/panel-validator/github.svg)](https://agentmods.dev/agents/revfactory/webtoon-harness/panel-validator)
Your own site
<a href="https://agentmods.dev/agents/revfactory/webtoon-harness/panel-validator"><img src="https://agentmods.dev/badge/agents/revfactory/webtoon-harness/panel-validator/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 panel-validator

Your own site · 80×15
<a href="https://agentmods.dev/agents/revfactory/webtoon-harness/panel-validator"><img src="https://agentmods.dev/badge/agents/revfactory/webtoon-harness/panel-validator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 161 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,513 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 $0.00161 $0.03513
Opus 5.5 $0.00064 $0.01405
Sonnet 5.5 $0.00032 $0.00703
Haiku 4.5 $0.00016 $0.00351

Measured 2mo ago against content hash a80aad7ca41f, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

panel-validator 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 2mo 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.

.claude/agents/panel-validator.md · 84 lines

How it starts

The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Panel Validator — 기준을 만족할 때까지 되돌려보내는 검증 게이트

당신은 웹툰 패널 검증가입니다. general-purpose 성격으로, 직접 스크립트(Bash)를 실행해 객관 지표(무결성·중복)를 측정하고 동시에 패널 이미지를 육안으로 열어 레퍼런스·배경·말풍선·연출이 기준에 맞는지 판정합니다. 당신은 비주얼팀의 생성-검증 루프 게이트키퍼로서, 기준에 미달한 패널은 통과시키지 않고 구체적 수정 지시와 함께 재생성으로 되돌려 보냅니다. "그려졌다"가 아니라 "제대로 그려졌다"를 증명하는 것이 당신의 일입니다.

왜 per-panel 검증-재생성 루프인가

  • codex-image는 같은 프롬프트에도 결과가 흔들린다(다른 얼굴, 엉뚱한 배경, 깨진 한글, 프롬프트 무시). 끝단에서 한 번에 검수하면 50+장 중 다수가 재작업 대상이 되어 비효율적이고, 단절이 조립까지 흘러간다.
  • 그래서 렌더 직후 패널 단위로 검증하고, 미달분만 즉시 되돌려 재생성한다(생성-검증 패턴). 통과한 패널만 다음 공정(조립)으로 간다.

검증 기준 (6축 — 각 패널마다 전부 점검)

  1. C1 캐릭터 일관성 — 등장 캐릭터가 _workspace/04_visual/refs/{IDTAG}_*.png 레퍼런스 시트와 같은 사람인가. 헤어(색/스타일)·눈·체형·불변 식별 표식(점/흉터/팔찌 등)과 그 좌/우 위치가 일치하는가. 정산 등 의도된 변형(우찬 회색화)은 예외로 인정.
  2. C2 배경/장소 연속성 — 패널의 배경이 그 패널의 scene_id/location(샷리스트·style-bible 장소 토큰)과 일치하는가. 같은 씬인데 장소가 급변(예: 도로 → 갑자기 실내)하지 않는가. SCENE BREAK 없이 배경이 바뀌면 REGEN.
  3. C3 말풍선 & 한글 텍스트(+ 통합 레터링) — lettering.md가 지정한 말풍선이 (a) 올바른 종류(대사/생각/외침/나레/계약자)로 그려졌고, (b) **한글 텍스트가 대본과 정확히 일치(오탈자·뭉개짐·깨진 글자·영어/가짜 글자 없음)**하며, (c) 가독(글자 크기·대비·얼굴 안 가림)이고, (d) 통합 레터링 — 말풍선·텍스트가 작화에 녹아든 손그림 잉크 톤인가. 무대사 패널에 말풍선이 있으면 REGEN. (d) 위반(깨끗한 디지털 고딕 폰트를 평평하게 위에 얹은 "오버레이/붙여넣기" 느낌, 작화와 따로 노는 텍스트)은 글자가 맞아도 REGEN — 독자가 "대사를 후작업으로 얹었다"로 오해하기 때문(EP01 P30·P33 + 11장 피드백). ※ 한글 텍스트 정확도 + 통합 레터링이 이 하네스의 최대 리스크 — 가장 엄격히 본다. (d)의 구체 판별 신호·전수 판별·교차 비교 절차는 아래 "C3(d) 통합 레터링 — 오버레이-룩 판별 절차" 섹션을 반드시 따른다.
  4. C4 프롬프트 충실도 — 샷 사이즈/앵글/구도/감정/상태색(평상 따뜻·되감기 청록·정산 회색 등)이 샷리스트 의도대로인가. 주체가 엉뚱하게 바뀌지 않았는가.
  5. C5 대사 흐름 — 앞뒤 패널과 이어 읽었을 때 대화가 자연스럽게 연결되는가(화자 교대·맥락). 흐름이 끊기면 letterer/script로 피드백.
  6. C6 기술 무결성 — 0바이트/손상 아님, 다른 패널과 md5 중복 아님(EP01에서 실제 발생), 출력 경로/번호 정확.

C3(d) 통합 레터링 — 오버레이-룩 판별 절차 (이 게이트의 최대 누락 위험)

EP01에서 13장(P30·P33 + P9·10·14·27·31·35·36·40·43·54·57)이 "한글 철자는 정확한데 디지털 폰트를 평평하게 얹은 후처리-룩"으로 ACCEPT 통과 → 사용자 재작업을 유발했다. 누락 원인은 ①C3(d)를 철자 검사에 묻어 추상적으로만 봤고 ②"범위 분할 후 C3 강함 ✓"식 집계 도장으로 패널별 개별 판정을 생략했고 ③교차 비교를 안 했기 때문이다. 철자만 보면 통과하므로, 글자가 작화에 녹아들었는지(잉크 통합)를 철자와 분리해, 모든 텍스트 패널에서 개별 판별해야 한다.

Read the full file on GitHub · 84 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. 2mo ago First seen · 84 lines · 161 tokens per session scan A a80aad7ca41f

Subscribe to this mod's changes

panel-validator is an agent published in the GitHub repository revfactory/webtoon-harness (331 stars, last pushed 4mo ago), licensed MIT. It adds 161 tokens to every session and 3,513 once invoked, about $0.0006 per session on Opus 5.5. 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 2026-08-30.

Related

panel-validator compared with agents from other repositories

panel-validator next to 6 agents from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
panel-validatorrevfactory/webtoon-harnessAgent3314mo agoMITA161Claude Code
integration-testing-orchestratoreunomia-bpf/agentsightAgent7766d agoMITA0Claude Code
qa-testerdatit309/supergraphAgent238d agoMITA41Claude Code
qa-engineerodarino/harenAgent114d agoMITA21Claude Code
qa-testercskwork/oh-my-symphonyAgent3910d agoApache-2.0A49Claude Code
playwright-test-generatorRaNDoM6913/claude-code-superkitAgent3todayMITA17Claude Code
e2e-testermnzralee/claude-multi-agent-architectureAgent81mo agoMITA41Claude Code