在架構決策、SDD 審查、技術選型、效能瓶頸分析時調度此 Agent
在功能實作、TDD 開發、代碼撰寫、Bug 修復時調度此 Agent
在需求定義、優先級決策、Sprint 規劃、Backlog 管理時調度此 Agent
在代碼審查、測試策略制定、Bug 重現、品質把關時調度此 Agent
在外部輸入處理、API 端點審查、配置安全變更、安全漏洞評估時調度此 Agent
Use when technical estimation, story sizing, ADR requirement assessment, or parallel dispatch grouping decisions are needed during Sprint Planning or Execution
Use when technical decisions are needed, architecture reviews, technology selection, ADR creation, or system design changes
Use when backlog grooming is needed, requirements change, new stories need RICE scoring, or backlog health maintenance is due
Use when browser-based testing or verification is needed — E2E tests, visual regression, form validation, responsive checks, or deployment smoke tests via agent-browser
Use when starting cruise mode, automated patrol, periodic issue scanning, or background monitoring — enables PO patrol + SRE inspection loops
Matches all tools
Hooks run on every tool call, not just specific ones
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
No model invocation
Executes directly as bash, bypassing the AI model
No model invocation
Executes directly as bash, bypassing the AI model
為你的 AI 開發工具注入 8 個專業角色,涵蓋 Discovery → Definition → Delivery 全產品生命週期。
| 新功能 | 說明 |
|---|---|
| Issue 建立安全封裝 | scripts/gh-issue-create.sh:強制 --body-file 模式,防止特殊字元截斷(紅線 #13 強化) |
| 禁用軟性字樣清單 | Architect / QA prompt 新增 8 條禁用詞表(考慮、適當、合理…),PO Round 1 自動自檢 |
| Agent fallback 驗證 | dispatch-with-fallback.sh + ADR-046:偵測 HTTP 429 / 5xx / Policy Refusal,自動切換模型 |
| ADR 歸檔政策 | audit-adr-references.sh(ACTIVE / DOCS-ONLY / DORMANT / SUPERSEDED 分類)+ ADR-009 已歸檔 |
| L-size 自動 Refinement | 排程模式下 L-size sprint-candidate 觀察期達標後,Cruise Patrol 自動觸發 Architect 拆解 |
/plugin marketplace add KCTW/shikigami
/plugin install shikigami
安裝後對 Claude 說:
幫我初始化 Shikigami— Scrum Master 會接手引導。
在 Claude Code 互動介面中輸入:
/plugin marketplace add KCTW/shikigami
/plugin install shikigami
安裝後重啟 Claude Code,在你的專案目錄說:
幫我初始化 Shikigami
Scrum Master 會建立 CLAUDE.md 與 docs/ 目錄結構,完成後即可開始。
Shikigami 提供 pre-commit hook,在 commit 前自動驗證版號一致性:
git config core.hooksPath .githooks
啟用後,每次 git commit 前會自動執行 bash scripts/validate-version.sh,版號不一致時阻止 commit。
| 平台 | 安裝指令 | 安裝指南 |
|---|---|---|
| OpenCode | symlink 適配策略 | docs/INSTALL_OPENCODE.md |
| Gemini CLI | gemini extensions install https://github.com/KCTW/shikigami | docs/INSTALL_GEMINI.md |
| Cursor | bash scripts/install-cursor.sh | docs/INSTALL_CURSOR.md |
你一個人開發,需求靠腦補、寫完的代碼沒人 review,架構決策靠直覺,安全問題等上線才發現。
Shikigami 注入 8 個角色,覆蓋產品從探索到上線的完整流程:
/performance-dashboard 一指令查看當日工作成果角色之間互相制衡,不是 8 個獨立助手。不需要記指令,用自然語言說你要做什麼,Scrum Master 自動調度。
我想加一個使用者登入功能,但不確定要用哪種方式
Shikigami 觸發 Discovery Phase:PO 產出 Product Brief(問題陳述 + 商業假設 + 驗證方法),Architect 評估技術可行性,PO 簽核後自動轉為 Backlog Issues。從探索到可執行 Story,全程結構化。
開始 Sprint / 實作 Story A
Developer subagent 依照 TDD 流程實作(Red → Green → Refactor),完成後 QA subagent 自動執行雙階段審查(Spec Compliance + Code Quality)。任何不符合 AC 的代碼都會在合併前被攔截。
開始 Sprint
Scrum Master → 觸發 sprint-planning
PO:從 Backlog 選取 Stories
Architect:技術估算
QA:AC 可測試性確認
PO:Sprint 文件已建立 → docs/sprints/sprint_N.md
實作 Story A
Developer subagent:
Red — 寫失敗測試
Green — 最小實作通過
Refactor — 重構優化
commit: "test: 新增功能測試"
commit: "feat: 實作核心邏輯"
QA subagent(Spec Review):所有 AC 通過
QA subagent(Code Quality):品質門禁通過
幫我分類一下 GitHub Issues
PO subagent:
gh issue list --search "no:label" → 找到 5 個未分類 issues
分類結果:
#12 登入失敗 | bug | 已要求補充重現步驟
#13 希望支援匯出 | feature-request | —
#14 怎麼安裝? | question | 已引導至 README
Labels 已自動套用
資料庫要用 PostgreSQL 還是 SQLite?
Architect subagent:
建立 ADR-002
選項 A:PostgreSQL — 擴展性強,維運成本高
選項 B:SQLite — 零配置,單檔案,不適合高併發
建議:採用 SQLite(MVP 階段,KISS 原則)
QA subagent(Decision Challenger):
為 PostgreSQL 辯護:未來遷移成本可能很高
結論:同意 SQLite,但建議抽象 DB 層以降低遷移風險
Architect:ADR-002 狀態 → Accepted
| 角色 | 職責 |
|---|---|
| Product Owner | 需求定義、優先級決策、Backlog 管理 |
| Architect | 架構決策、SDD 撰寫、技術選型 |
| Developer | 功能實作、TDD 開發、Bug 修復 |
| QA Engineer | 代碼審查、測試策略、品質把關 |
| Security Engineer | 安全掃描、漏洞評估、OWASP 檢查 |
| SRE Engineer | 部署檢查、監控配置、環境管理 |
| UI/UX Designer | 設計系統維護、Figma 原型製作、視覺品質審查 |
| Stakeholder | 最終仲裁、打破僵局 |
重點:它們互相制衡,不是 8 個獨立助手。
Discovery(產品探索)
| Skill | 說明 |
|---|---|
| discovery-phase | Product Discovery 獨立入口、Product Brief 標準化格式、假設外顯化、PO 確認關卡 |
Definition(需求定義與 Sprint 管理)
| Skill | 說明 |
|---|---|
| scrum-master | 自動調度 Agent Scrum Team 的角色分工與 Sprint 流程 |
| sprint-planning | 啟動新 Sprint、從 Backlog 選取 Stories、規劃 Sprint 目標 |
| backlog-management | Backlog 梳理、需求變更管理 |
| escalation | 團隊衝突無法解決、重大產品轉向、升級鏈啟動 |
Delivery(開發與交付)
npx claudepluginhub kctw-dev/shikigami --plugin shikigamiComplete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, and rules evolved over 10+ months of intensive daily use
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Comprehensive .NET development skills for modern C#, ASP.NET, MAUI, Blazor, Aspire, EF Core, Native AOT, testing, security, performance optimization, CI/CD, and cloud-native applications
Complete creative writing suite with 10 specialized agents covering the full writing process: research gathering, character development, story architecture, world-building, dialogue coaching, editing/review, outlining, content strategy, believability auditing, and prose style/voice analysis. Includes genre-specific guides, templates, and quality checklists.
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.