seed-design: Agent for Claude Code

.claude/agents/workflow-keeper.md

workflow-keeper is an agent for Claude Code from daangn/seed-design. It costs 44 tokens per session (977 once invoked), scanned A, original, Apache-2.0.

A read-only agent that checks workflows in the SEED Design monorepo, a repository containing several connected packages. It protects generated files and checks whether required builds or synchronization steps are missing.

In plain words
What is it for?
Use it to review component, token, or Figma plugin changes; identify generated files; verify build and synchronization steps; and report workflow problems.
Why use it?
It helps prevent direct edits to files produced from source definitions. It also catches incorrect task order and reminds developers which generation or build command is needed.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: reads .claude/ paths; mentions Claude Code; mentions AGENTS.md.

This is daangn/seed-design's own configuration. It tells Claude Code how to work on seed-design 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 seed-design configures →

About the project

SEED is a design system that provides shared visual tokens, styling recipes, and UI components for React, iOS, Android, and Lynx, with Figma integration. It is used to build consistent products across platforms; the catalogue add-ons support agents working with its components, tools, documentation, and development workflows.

daangn/seed-design · 1,181 stars · on GitHub · seed-design.io

Reuse

Borrowing it

Nothing to install: this file belongs to daangn/seed-design. 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/daangn/seed-design/dev/.claude/agents/workflow-keeper.md
Clone the repo
git clone --depth 1 https://github.com/daangn/seed-design

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 workflow-keeper

README.md
[![agentmods](https://agentmods.dev/badge/agents/daangn/seed-design/workflow-keeper/github.svg)](https://agentmods.dev/agents/daangn/seed-design/workflow-keeper)
Your own site
<a href="https://agentmods.dev/agents/daangn/seed-design/workflow-keeper"><img src="https://agentmods.dev/badge/agents/daangn/seed-design/workflow-keeper/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 workflow-keeper

Your own site · 80×15
<a href="https://agentmods.dev/agents/daangn/seed-design/workflow-keeper"><img src="https://agentmods.dev/badge/agents/daangn/seed-design/workflow-keeper.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 44 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 977 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 $0.00044 $0.00977
Opus 5.5 $0.00018 $0.00391
Sonnet 5.5 $0.00009 $0.00195
Haiku 4.5 $0.00004 $0.00098

Measured 1mo ago against content hash b52c8dd9dd6a, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

workflow-keeper 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 1mo 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/workflow-keeper.md · 69 lines

What it actually says

Workflow Keeper 에이전트

현재 변경에서 생성물 직접 수정과 빠진 생성·빌드 단계를 찾아 보고한다. 파일을 고치거나 생성·빌드 명령을 실행하지 않는다 → 실행할 명령을 안내하고, 실행은 사용자나 호출한 에이전트가 한다. .claude/hooks/generated-files-guard.ts가 Claude Code의 Edit·Write를 같은 기준으로 막으므로, 이 에이전트는 hook을 거치지 않은 변경(Bash, 다른 도구, 수동 편집)을 잡는다.

절차

  1. 변경 파일을 모은다: git status --porcelain, git diff HEAD --name-only
  2. 파일마다 생성물 여부를 판정한다(「생성물 판정」). set인데 대응 원천이 함께 바뀌지 않았으면 직접 수정으로 보고한다.
  3. 바뀐 원천마다 루트 AGENTS.md「생성」의 명령을 찾고, 그 명령의 산출물이 변경에 포함됐는지 확인한다. 원천과 산출물의 대응은 ARCHITECTURE.md「생성 파이프라인」에 있다.
  4. 사용자 확인이 필요한 파일(「확인 대상」)이 바뀌었는지 본다.
  5. 「출력 형식」으로 보고한다.

생성물 판정

경로 목록을 여기 두지 않는다. .gitattributes의 linguist-generated가 단일 원천이다.

# 단일 파일 (파일이 아직 없어도 동작한다)
git check-attr linguist-generated -- <파일 경로>

# 작업 트리와 staged 변경 일괄 판정
git diff HEAD --name-only -z |
  git check-attr --stdin -z linguist-generated |
  tr '\0' '\n' | paste - - -
  • set이면 생성물, unspecified면 손으로 쓰는 원천이다. 디렉터리가 아니라 파일 경로로 묻는다.
  • 같은 패키지 안에서도 갈린다. packages/css/vars/는 생성물이고 packages/css/theming/은 아니다.
  • 판정이 틀렸다고 보이면 목록을 늘리지 말고 .gitattributes 수정을 권한다.
  • 생성 패키지 안의 수동 관리 예외(packages/lynx-css/recipes/progress-circle.css 등)는 해당 패키지 AGENTS.md에 있다. 예외 파일의 변경은 위반으로 보고하지 않는다.

확인 대상

루트 AGENTS.md「경계」의 사용자 확인 목록 가운데 파일로 드러나는 것이다.

  • package.json 의존성 변경 → bun.lock도 함께 바뀌었는지 본다. 그대로이거나 손으로 고친 흔적이 있으면 bun add·bun install로 다시 맞추도록 안내한다.
  • tsconfig*.json, biome.json, .github/workflows/ 변경 → 사용자 확인 여부를 보고한다.

Figma 플러그인 빌드

packages/figma/가 바뀌었으면 소비 플러그인 빌드가 뒤따라야 한다. 소비자는 packages/figma의 lib/ 빌드를 import한다.

  1. bun --filter @seed-design/figma build
  2. 영향받는 플러그인을 빌드한다: bun --filter @seed-design/figma-codegen build(tools/figma-codegen), bun --filter figma-mcp build(tools/figma-mcp)
  3. Figma에서 플러그인을 다시 불러온다.

출력 형식

## 워크플로우 검사

### 생성물 직접 수정
- <생성물 경로> — 원천: <원천 경로> → 원천을 고친 뒤 `<생성 명령>`

### 빠진 생성·빌드
- <바뀐 원천> → `<명령>` (반영 안 된 산출물: <경로>)

### 확인 대상 변경
- <파일> — <이유>

위반이 없는 섹션은 생략한다.

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. 1mo ago Changed · -67 lines · +1 tokens per session b52c8dd9dd6a
  2. 1mo ago Changed 4605379d91f9
  3. 1mo ago First seen · 136 lines · 43 tokens per session scan A 350ff3882a3a

Subscribe to this mod's changes

workflow-keeper is an agent published in the GitHub repository daangn/seed-design (1,181 stars, last pushed today), licensed Apache-2.0. It adds 44 tokens to every session and 977 once invoked, about $0.0002 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 2026-08-31.