chiyum/claude-code-config
Claude Code 完整配置範本：多 Agent 協作（architect/reviewer/QA/PM）、自訂 Skill、產品驗收體系、Slack Hook 整合
ARCHETYPE
YOLO Cowboy
Permissions? Never heard of 'em. Ships at the speed of `--dangerously-skip-permissions`.
Copy this rig
# review before running: this installs third-party code $ claude mcp add playwright -- npx @anthropic-ai/mcp-playwright $ npx degit chiyum/claude-code-config/agents ./rig-claude-code-config/agents $ npx degit chiyum/claude-code-config/skills ./rig-claude-code-config/skills $ npx degit chiyum/claude-code-config/hooks ./rig-claude-code-config/hooks
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ claude mcp add playwright -- npx @anthropic-ai/mcp-playwright $ npx degit chiyum/claude-code-config/skills/dev .claude/skills/dev $ npx degit chiyum/claude-code-config/skills/discover .claude/skills/discover $ npx degit chiyum/claude-code-config/skills/merge-prod .claude/skills/merge-prod $ npx degit chiyum/claude-code-config/skills/retro .claude/skills/retro $ npx degit chiyum/claude-code-config/skills/spec-check .claude/skills/spec-check $ npx degit chiyum/claude-code-config/skills/test-line-message-integrity .claude/skills/test-line-message-integrity $ npx degit chiyum/claude-code-config/skills/test-web-message-integrity .claude/skills/test-web-message-integrity $ npx degit chiyum/claude-code-config/skills/verify-ocr-version .claude/skills/verify-ocr-version
$ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/architect.md $ curl -fsSL --create-dirs -o .claude/agents/design-reviewer.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/design-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/pm.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/pm.md $ curl -fsSL --create-dirs -o .claude/agents/qa.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/qa.md $ curl -fsSL --create-dirs -o .claude/agents/reviewer.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/security-auditor.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/security-auditor.md $ curl -fsSL --create-dirs -o .claude/agents/ui-designer.md https://raw.githubusercontent.com/chiyum/claude-code-config/main/agents/ui-designer.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 7,204 rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(rm -rf *)",
"Read(**/*.pem)",
"Bash(rm -rf /)",
"Bash(git push --force:*)",
"Bash(sudo *)",
"Read(.env)",
"Bash(rm -rf /*)",
"Read(./.env.*)",
"Read(~/.aws/**)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(sudo:*)",
"Bash(git reset --hard*)",
"Bash(git reset --hard:*)",
"Read(.env.*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(rm:*)",
"Bash(git rebase *)",
"Bash(wget *)",
"Bash(npm publish:*)",
"Bash(git commit *)",
"Bash(gh pr merge *)"
]
}
} MCP servers (1)
| server | source | est. tokens |
|---|---|---|
| @anthropic-ai/mcp-playwright · "playwright" | npm | 2.5k |
Skills (8)
devdiscovermerge-prodretrospec-checktest-line-message-integritytest-web-message-integrityverify-ocr-version
Subagents (7)
| architect model: opus | 資深工程師兼架構師。負責所有「會修改程式碼」的工作。收到新需求時先提出三種技術方案（含優劣、潛在風險、回滾難度、架構圖、測試策略），選定方案後再實作。觸發時機：使用者提出新需求、bug fix、refactor、新增功能、跨模組改動等任何需要寫/改 code 的工作。 |
| design-reviewer model: opus | 視覺與互動品質審查員（反方設計師）。在 architect 完成 UI/版面改動、reviewer 通過後，用 Playwright 對頁面做三視口截圖審查，對照 ui-designer 的設計規格批判版面構圖、間距節奏、字體階層、色彩、動效與狀態完整度，裁決 PASS 或退修並給出數值級的具體修改清單。只審查不改 code。觸發時機：與 ui-designer 成對——本任務開過正方（設計意圖任 |
| pm model: opus | 通用產品經理（Product Manager）。職責是閱讀並理解規格書，當開發完成某個功能時，比對實際運行結果與規格書的差異，用 Playwright MCP 操作瀏覽器走完整使用者流程驗收。回報「符合規格 / 不符合 / 缺漏 / 多做」並列出證據（截圖、URL、實際 vs 預期對照）。不能直接改 code，只能讀取與驗證。本 agent 與產品解耦：每次驗收前先從 `~/.claude/pro |
| qa model: opus | 通用 QA 測試員（產品無關）。在 commit 後進行本地與線上 dev 測試。使用 Playwright MCP 測前端 UI、用 curl/API 腳本測後端。每次任務前先從 ~/.claude/products/INDEX.md 載入對應產品配置（環境、帳號、URL、部署方式都以配置為準）。驗收證據一律落地為檔案。回報測試結果，不下放行決策。 |
| reviewer model: opus | 在程式碼修改完成後、commit 前進行審查。檢查程式碼品質、架構合理性、是否符合需求、潛在隱患（安全、效能、邊界條件、資料一致性）。回報通過或具體問題清單。 |
| security-auditor model: opus | 資安審查專員（基礎設施 / 伺服器面）。在「架設或變更伺服器 / 部署環境 / 對外服務」的任務完成後，對該環境做防禦式資安審查——網路暴露面、認證與存取控制、機敏資料與金鑰、傳輸與儲存加密、容器 / 反向代理 / DB / 防火牆設定、日誌與稽核。以「證明這台伺服器可被攻破」為預設立場，逐項給 PASS / FAIL + 證據 + 修法。只審查與唯讀探測，絕不改設定、絕不做破壞性或攻擊性操作。觸 |
| ui-designer model: opus | 視覺設計師（正方）。在設計/切版/視覺類任務動 code 之前，產出可直接實作的「設計規格」——美學方向、token 對映、版面構圖、數值級動效 spec、三態設計；大型視覺任務先產三個 direction 的靜態 HTML 預覽供使用者挑選。產出規格與預覽檔，不改專案 code。觸發時機：①使用者指令含設計/切版/視覺/風格/版面意圖的任務（依 DESIGN_FLOW 觸發判準）②功能任務中要從 |
Hooks (1)
| event | matcher | runs |
|---|---|---|
| Stop | * | $HOME/.claude/hooks/slack-notify.sh |
Plugins (5)
gopls-lsp@claude-plugins-officialplaywright@claude-plugins-officialtelegram@claude-plugins-officialfrontend-design@claude-plugins-officialgsap-skills@gsap-skills
Permissions
deny (0)
—
ask (0)
—
allow (46)
Edit
Write
Bash(*)
mcp__playwright__browser_navigate
mcp__playwright__browser_snapshot
mcp__playwright__browser_click
mcp__playwright__browser_type
mcp__playwright__browser_fill_form
mcp__playwright__browser_take_screenshot
mcp__playwright__browser_press_key
mcp__playwright__browser_hover
mcp__playwright__browser_select_option
mcp__playwright__browser_tabs
mcp__playwright__browser_navigate_back
mcp__playwright__browser_evaluate
mcp__playwright__browser_console_messages
mcp__playwright__browser_network_requests
mcp__playwright__browser_file_upload
mcp__playwright__browser_handle_dialog
mcp__playwright__browser_drag
mcp__playwright__browser_wait_for
mcp__playwright__browser_resize
mcp__playwright__browser_run_code
mcp__playwright__browser_close
mcp__playwright__browser_install
mcp__plugin_playwright_playwright__browser_navigate
mcp__plugin_playwright_playwright__browser_snapshot
mcp__plugin_playwright_playwright__browser_click
mcp__plugin_playwright_playwright__browser_type
mcp__plugin_playwright_playwright__browser_fill_form
mcp__plugin_playwright_playwright__browser_take_screenshot
mcp__plugin_playwright_playwright__browser_press_key
mcp__plugin_playwright_playwright__browser_hover
mcp__plugin_playwright_playwright__browser_select_option
mcp__plugin_playwright_playwright__browser_tabs
mcp__plugin_playwright_playwright__browser_navigate_back
mcp__plugin_playwright_playwright__browser_evaluate
mcp__plugin_playwright_playwright__browser_console_messages
mcp__plugin_playwright_playwright__browser_network_requests
mcp__plugin_playwright_playwright__browser_file_upload
mcp__plugin_playwright_playwright__browser_handle_dialog
mcp__plugin_playwright_playwright__browser_drag
mcp__plugin_playwright_playwright__browser_wait_for
mcp__plugin_playwright_playwright__browser_resize
mcp__plugin_playwright_playwright__browser_run_code
mcp__plugin_playwright_playwright__browser_close
Similar rigs
itshig/AGENTS-template
Drop-in AGENTS.md template for Claude Code, Cursor, Codex, and other AI agents — with .agents/ for rules, personas, and workflows.
Orchestrator 3.0k tok ·
eugeek/claude-config-tdd
Claude Code config with TDD workflow, plan -> tests -> red -> implementation -> green -> review -> shipment
Orchestrator 5.8k tok ·
IntuitivePhella/AI-PhellOS
Portable AI development workflow template and CLI for spec-driven, test-aware software delivery with Codex, Claude Code, and other coding agents.
Orchestrator 3.1k tok ·
xhulz/nina
Harness orchestration for Claude Code: composes a gated multi-agent pipeline into projects from versioned layers, holds its loop caps with hooks, measures it from transcripts, and turns recurring mistakes into rules.
Orchestrator 9.3k tok ·