Fable Commander

Fable 指揮官工作流 — Fable 只做規劃/審查/決策，研究與執行派給使用者選的模型（Opus/Sonnet/Haiku subagent）。流程：Fable 產 plan → 問使用者選研究模型 → /workflow 查資料+列驗收標準 → Fable review 修 plan → 問使用者選執行模型 → maker/verifier 分離的執行 loop（驗證硬條件=客觀正確性，禁止以風格/觀點打回）。觸發詞：指揮官模式、fable commander、指揮官流程、用指揮官跑、Fable 當指揮官、commander 這個題目。

35 stars 12 forksProductivity & Notes

Install

skills CLI (Claude Code, Cursor, Codex and more)

npx skills add DennisWei9898/fable-commander --skill fable-commander

Manual install

git clone --depth 1 https://github.com/DennisWei9898/fable-commander.git /tmp/fable-commander
mkdir -p ~/.claude/skills
cp -r /tmp/fable-commander/. ~/.claude/skills/fable-commander

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# fable-commander · Fable 指揮官工作流

你（Fable）是**指揮官**：負責規劃、審查、整合、把關決策點。查資料與寫程式這類大量 token 的工作，派給**使用者選定的模型**用 subagent / Workflow 跑。核心信念來自 Addy Osmani 的 Loop Engineering：

> "The most useful structural thing in a loop, by far, is splitting the one who writes from the one who checks. The model that wrote the code is way too nice grading its own homework."

## 模型切換的實作方式（重要，先讀懂）

Claude Code 的 main loop 模型**無法由 Claude 自行切換**（`/model` 是使用者指令）。因此本 skill 的「切模型」一律用：
- `Agent` / `Workflow` 的 `model` 參數（`opus` / `sonnet` / `haiku`）— 自動、不需使用者動手 ← **預設**
- 若使用者明確想把 main loop 整個換掉（例如長執行期省成本），提示他自己輸入 `/model`，並把 handoff 所需的一切寫進 state file 再交棒

## Advisor 模式：本 skill 的輕量替身（官方 harness 功能，非工作流）

`/advisor <model>` 是 Claude Code 官方（實驗性）功能：主模型跑便宜的（如 Sonnet），Claude 在關鍵時刻（提交方案前／錯誤重複／宣告完成前）**自動諮詢**更強的 advisor（如 Opus）。設定方式 `/advisor opus`、`advisorModel` settings 欄位、或 `--advisor` 旗標；`/advisor off` 關閉。需 Claude Code v2.1.98+，僅 Anthropic API（Bedrock/Vertex/Foundry 不支援）。mid-session 即生效、不需重開 session（官方文件明說中途啟用停用不使 prompt cache 失效）。

**何時用它取代本 skill**：官方文件原文是 advisor 適合「**長、多步驟**任務，方案品質決定成敗」，短任務「加值較低」，官方建議短任務直接換主模型而非掛 advisor——所以正確的折衷地帶，是 Phase −1 三判準裡的「**驗證難以客觀化**」或「**token 預算不夠**」這兩類、但任務本身仍偏長偏多步驟；純粹「任務小」不是 advisor 的甜蜜點，照 Phase −1 直接做即可。

**為何它不能取代完整 loop**：advisor 是給「同一個執行者」的第二意見，執行者通常照建議走——這跟本 skill 的 maker/verifier 分離（獨立、有罪推定、跨模型的檢查者）**方向相反**；官方 blog 明說 advisor strategy 是「反過來的 orchestrator」。所以 advisor 沒有獨立驗證、沒有 gate/停止紀律、沒有證據要求。防「誤報完成」的護欄只有完整 loop 才有。

Read the full skill on GitHub

Details

Skill name: fable-commander
Repository: DennisWei9898/fable-commander
License: MIT
Repo updated: 12 July 2026

FAQ

What does the Fable Commander skill do?

How do I install the Fable Commander skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add DennisWei9898/fable-commander --skill fable-commander. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Fable Commander popular?

It lives in DennisWei9898/fable-commander, which has 35 GitHub stars and 12 forks.

Related skills

Slack Agent

vercel-labs/slack-agent-skill

Use when building Slack agents/bots with eve (Vercel's filesystem-first agent framework), @vercel/connect, or eve/channels/slack. Covers defineAgent/defineTool patterns, Vercel Connect credential brokering, Slack channel setup, testing requirements, and quality standards.

Productivity & Notes

Trdizin

saidsurucu/trdizin-skill

Use when the user wants to search TR Dizin (trdizin.gov.tr — TÜBİTAK ULAKBİM national academic index) for publications, journals, authors, or institutions; do field-scoped advanced search (title/author/abstract/journal/doi/year/…); apply facet filters (access type, year, language, subject, document type); read a publication's references; or extract a publication PDF as text. Uses the open public JSON API over plain HTTPS — no login, no CAPTCHA, no API key.

Productivity & Notes

Mastering Langgraph

SpillwaveSolutions/mastering-langgraph-agent-skill

Build stateful AI agents and agentic workflows with LangGraph in Python. Covers tool-using agents with LLM-tool loops, branching workflows, conversation memory, human-in-the-loop oversight, and production monitoring. Use when - (1) building agents that use tools and loop until task complete, (2) creating multi-step workflows with conditional branches, (3) adding persistence/memory across turns with checkpointers, (4) implementing human approval with interrupt(), (5) debugging via time-travel or LangSmith. Covers StateGraph, nodes, edges, add_conditional_edges, MessagesState, thread_id, Command

Productivity & Notes

Workflow Design Bible

preangelleo/workflow-design-bible

Generate a complete "constitution + documentation system" for a new autonomous, agent-run project — a content channel, an ebook press, an SEO tool-site, a web product, a casual game, or anything that should run itself with minimal human babysitting. Use when starting a brand-new self-running project and you want a CEO-orchestrated architecture (main agent → sub-agents → skills → CLI/MCP) with a named document system, a session lifecycle (start → work → finalize), and a growing identity/soul — scaffolded from a short structured interview. Triggers - "start a new autonomous project", "scaffold a

Productivity & Notes

Wechat Clawbot Notify

plustar35/wechat-clawbot-notify

Send notification messages to the user's WeChat via ClawBot (iLink API). Use when automation tasks complete and need to notify the user, or when the user asks to send a message to their WeChat. Triggers on: 'notify me on WeChat', 'send to my WeChat', 'push result to WeChat', '发微信通知', '推送到微信', '通知我'.

Productivity & Notes

Memory Bank

fockus/skill-memory-bank

Agent-agnostic long-term project memory through `.memory-bank/` + RULES (TDD/SOLID/Clean Architecture/FSD/Mobile) + dev-toolkit commands. Use when working in a project with a `.memory-bank/` directory or when the user explicitly asks for memory-bank workflow, code rules, or dev-toolkit commands (`/mb`, plan, spec, verify, done; «память проекта», «контекст проекта», «план», «продолжай по плану»).

Productivity & NotesIn 1 list

Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.