Token导航 LogoToken导航

Agent Skill · 待分类

crucible-planner(坩埚规划师)

crucible-planner 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

crucible-planner 是待分类类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在待分类类的 11,209 个 Skill 里排第 9,293。

总安装 247周安装 10GitHub Stars 48下载 78核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:crucible-planner(坩埚规划师)
来源仓库:https://github.com/forsonny/the-crucible-writing-system-for-claude
仓库路径:skills/crucible-planner
安装命令:
npx skills add https://github.com/forsonny/the-crucible-writing-system-for-claude --skill crucible-planner
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/forsonny/the-crucible-writing-system-for-claude --skill crucible-planner

这个 Skill 是做什么的

这个 Skill 的分类还没最终确定,用法以来源仓库的说明为准。

crucible-planner 通过多选问答引导构建七类关联规划文档,支撑大型 fantasy 小说筹备。适用于世界观搭建、人物关系网络与剧情伏笔设计的系统性前期准备工作。涵盖哲学核心、线索地图、危机蓝图与反派设定等模块,强调内在逻辑自洽。

启动前必读全部参考资料,否则可能生成碎片化内容难以整合成 cohesive 故事线。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/crucible-planner

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/forsonny/the-crucible-writing-system-for-claud…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

crucible-planner 怎么安装?

在终端执行 npx skills add https://github.com/forsonny/the-crucible-writing-system-for-claude --skill…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

crucible-planner 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

crucible-planner 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

crucible-planner 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

crucible-planner 用的人多吗?

公开数据是累计安装 247 次,近一周安装 10 次,下载 78 次,GitHub 48 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

crucible-planner 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理待分类类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

待分类类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
infsh-cli

infsh CLI · Codex、Claude

4,374,576382权限未标注未明确未收录
ai-video-generation

AI 视频生成 · Codex、Claude

4,313,376382调用外部服务未明确有提醒
lark-openapi-explorer

云雀 OpenAPI 浏览器 · Codex、Claude

2,064,1209,011需要联网未明确有提醒
lark-workflow-standup-report

Lark 工作流程站立报告 · Codex、Claude

2,041,6328,978需要联网未明确通过
use-my-browser

使用我的浏览器 · Codex、Claude

1,848,49140操作浏览器未明确未通过
crucible-planner

坩埚规划师 · 本页

24748只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Crucible Planner

Interactive planning system for epic fantasy novels using the Crucible Structure.

Overview

Starting from a simple premise, guide users through multi-choice questions to build seven interconnected planning document categories (14 files total):

  1. Crucible Thesis — philosophical core
  2. Strand Maps — Quest, Fire, Constellation (3 separate maps)
  3. Forge Point Blueprints — five convergence crises
  4. Dark Mirror Profile — antagonist design
  5. Constellation Bible — character relationships
  6. Mercy Ledger — mercy/payoff tracking
  7. World Forge — world-building

Before Starting

Always read these references:

  • references/crucible-structure.md (the 36-beat structure)
  • references/question-sequences.md (complete question flows)

Workflow

Phase 1: INTAKE → Accept premise, initialize state, confirm scope
Phase 2: QUESTIONING → 9 document cycles with multi-choice questions
Phase 3: COMPILATION → Generate documents, present package

Phase 1: Intake

Accept Premise

Extract from user input:

  • Core concept (1-2 sentences)
  • Protagonist sketch
  • Central conflict hint

If premise too vague, use AskUserQuestion:

{
  "questions": [
    {
      "header": "Protagonist",
      "question": "Who is your protagonist?",
      "options": [
        {"label": "Reluctant chosen one", "description": "A chosen one who doesn't want the role"},
        {"label": "Ordinary hero", "description": "An ordinary person thrust into extraordinary circumstances"},
        {"label": "Fallen power", "description": "A powerful figure who's lost everything"},
        {"label": "Seeking redemption", "description": "A morally gray character seeking redemption"}
      ],
      "multiSelect": false
    }
  ]
}

Initialize State

python scripts/init_project.py "./crucible-project" "Title" "Premise"

Confirm Scope

Use AskUserQuestion to confirm scope:

{
  "questions": [
    {
      "header": "Novel Length",
      "question": "What is your target novel length?",
      "options": [
        {"label": "Standard", "description": "100-150K words, ~20-25 chapters"},
        {"label": "Epic", "description": "150-250K words, ~25-35 chapters"},
        {"label": "Extended/Series", "description": "250K+ words or multi-book series, 35+ chapters"}
      ],
      "multiSelect": false
    },
    {
      "header": "Complexity",
      "question": "What is your narrative complexity?",
      "options": [
        {"label": "Single POV", "description": "Single protagonist focus"},
        {"label": "Dual POV", "description": "Two protagonists sharing the story"},
        {"label": "Ensemble", "description": "Multiple POVs (3-5 characters)"}
      ],
      "multiSelect": false
    }
  ]
}

Phase 2: Document Generation

Questioning Rules

  1. ALWAYS use AskUserQuestion tool for all questions (provides interactive UI)
  2. Max 4 options per question (tool limit) + "Other" is automatic
  3. Max 4 questions per AskUserQuestion call
  4. Reference previous answers to build coherence
  5. Save state IMMEDIATELY after EACH answer (see Answer Persistence below)
  6. Verify before moving to next document

Answer Persistence Workflow

CRITICAL: Save every answer immediately after receiving it with FULL CONTEXT. This ensures progress is preserved if the session fails and answers can be understood later.

Save Command Format

Each answer is saved with the question text, selected answer, and description:

python scripts/save_state.py "<project_path>" --answer "<document>" "<state_key>" "<question_text>" "<answer>" "<description>"

Parameters:

  1. document - The document key (e.g., "doc1_crucible_thesis")
  2. state_key - The field name (e.g., "burden_type")
  3. question_text - The full question that was asked
  4. answer - The selected option label
  5. description - The description of the selected option
Step-by-Step Workflow
  1. Ask the question using AskUserQuestion: {"questions": [{"header": "Burden", "question": "What form does the external burden take in your story?", "options": [{"label": "Physical object", "description": "An artifact that must be destroyed or protected"}, {"label": "Person to save", "description": "Someone who must be rescued or kept alive"}], "multiSelect": false}]}
  2. User selects: "Physical object"
  3. IMMEDIATELY save with full context: python scripts/save_state.py "./crucible-project" --answer "doc1_crucible_thesis" "burden_type" "What form does the external burden take in your story?" "Physical object" "An artifact that must be destroyed or protected"
  4. Verify save succeeded (output shows question and answer)
  5. Ask next question
What Gets Saved

Each answer is stored as:

{
  "question": "What form does the external burden take in your story?",
  "answer": "Physical object",
  "description": "An artifact that must be destroyed or protected"
}

This provides full context when resuming or reviewing progress.

Example Save Commands

Document 1 (Crucible Thesis):

python scripts/save_state.py "./project" --answer "doc1_crucible_thesis" "burden_type" "What form does the burden take?" "Physical object" "An artifact to destroy or protect"
python scripts/save_state.py "./project" --answer "doc1_crucible_thesis" "fire_type" "What is the Fire's nature?" "Magical ability" "A power that corrupts with use"

Nested Documents (Forge Points, Mercies):

python scripts/save_state.py "./project" --answer "doc5_forge_points.fp0_ignition" "quest_crisis" "What Quest crisis emerges?" "Artifact stolen" "Enemy takes the burden"
python scripts/save_state.py "./project" --answer "doc8_mercy_ledger.mercy_1" "recipient" "Who receives mercy?" "Enemy soldier" "A combatant who surrendered"
On Document Completion

After the verification question is confirmed:

python scripts/save_state.py "./crucible-project" --complete <doc_num>

This marks the document complete and advances progress to the next document.

Error Handling

If a save fails:

  1. Retry once
  2. If still failing, inform the user but continue (answer is in context)
  3. User can manually recover using /crucible-continue

Question Format

CRITICAL: Use the AskUserQuestion tool, NOT plain text options.

Example AskUserQuestion call:

{
  "questions": [
    {
      "header": "Burden",
      "question": "What form does the external burden take in your story?",
      "options": [
        {"label": "Physical object", "description": "An artifact that must be destroyed or protected"},
        {"label": "Person to save", "description": "Someone who must be rescued or kept alive"},
        {"label": "Knowledge/truth", "description": "Information that must be revealed or protected"},
        {"label": "Mission/quest", "description": "A task that must be completed"}
      ],
      "multiSelect": false
    }
  ]
}

Header examples (max 12 chars): "Burden", "Fire Type", "Bond", "Antagonist", "Theme", "Sacrifice"

For verification questions, use multiSelect: true to allow checking multiple items.

Document Sequence

See references/question-sequences.md for complete question banks.

Document 1: Crucible Thesis (10 questions)

  • The Burden (external mission)
  • The Fire (internal power/curse)
  • Core Constellation Bond
  • Dark Mirror Connection
  • Forging Question
  • Antagonist's Truth
  • The Surrender
  • Theme
  • Blade's Purpose
  • Verification

Document 2: Quest Strand Map (7 questions)

  • Burden's Origin
  • Why This Protagonist
  • Antagonist's Stake
  • Quest Escalation
  • Impossible Requirement
  • Resolution Method
  • Verification

Document 3: Fire Strand Map (7 questions)

  • Fire Manifestation
  • The Danger
  • Cost of Use
  • Mastery Path
  • Hardening Phase
  • Mastery Moment
  • Verification

Document 4: Constellation Strand Map (7 questions)

  • Faithful Companion
  • The Sacrifice
  • Betrayal Source
  • Expansion
  • Bond That Saves
  • Constellation Fate
  • Verification

Document 5: Forge Point Blueprints (5 × 4 questions) For each Forge Point (Ignition, First, Second, Third, Apex):

  • Quest Crisis
  • Fire Crisis
  • Constellation Crisis
  • What is Sacrificed

Document 6: Dark Mirror Profile (9 questions)

  • Origin Parallel
  • The Divergence
  • Antagonist's Want
  • Compelling Offer
  • Why Tempting
  • Hidden Cost
  • Defeat Method
  • Antagonist's End
  • Verification

Document 7: Constellation Bible (12 questions)

  • Protagonist Profile
  • Faithful Companion details
  • Core Bond Character details
  • Mentor/Catalyst
  • Sacrifice Character
  • Additional cast
  • Verification

Document 8: Mercy Ledger (4 × 4 questions) For each of 4 mercies:

  • Recipient
  • Merciful Act
  • Immediate Cost
  • Later Payoff

Document 9: World Forge (9 questions)

  • World's Wound
  • Power System Source
  • Power Limitations
  • Previous Wielders
  • Key Locations
  • World-Protagonist Mirror
  • Magic Rules
  • Timeline Framework
  • Verification

Phase 2 Complete

After completing Document 9 verification, Phase 2 is complete. All foundational planning documents have been created.

Proceed to Phase 3 to compile the documents into the final planning structure by running compile_documents.py.


Phase 3: Compilation

Generate Documents

python scripts/compile_documents.py "./crucible-project"

Creates:

planning/
├── crucible-thesis.md
├── strand-maps/
│   ├── quest-strand.md
│   ├── fire-strand.md
│   └── constellation-strand.md
├── forge-points/
│   ├── fp0-ignition.md
│   ├── fp1-first-crucible.md
│   ├── fp2-second-crucible.md
│   ├── fp3-third-crucible.md
│   └── apex-willed-surrender.md
├── dark-mirror-profile.md
├── constellation-bible.md
├── mercy-ledger.md
├── world-forge.md
└── crucible-summary.md

Present to User

✅ **Crucible Planning Complete!**

📄 [View Crucible Thesis](computer:///path/planning/crucible-thesis.md)
📄 [View Strand Maps](computer:///path/planning/strand-maps/)
📄 [View Forge Points](computer:///path/planning/forge-points/)
📄 [View Dark Mirror Profile](computer:///path/planning/dark-mirror-profile.md)
📄 [View Constellation Bible](computer:///path/planning/constellation-bible.md)
📄 [View Mercy Ledger](computer:///path/planning/mercy-ledger.md)
📄 [View World Forge](computer:///path/planning/world-forge.md)
📋 [View Quick Reference](computer:///path/planning/crucible-summary.md)

After presenting the document links, use AskUserQuestion for next steps:

{ "questions": [ { "header": "Next Steps", "question": "What would you like to do next?", "options": [ {"label": "Review documents", "description": "Review and adjust any planning document"}, {"label": "Begin outline", "description": "Start creating chapter outlines from your plan"}, {"label": "Start drafting", "description": "Jump straight into writing prose"} ], "multiSelect": false } ] }


## State Management

Save after every question cluster:

python scripts/save_state.py "./crucible-project"


Resume interrupted session:

python scripts/load_state.py "./crucible-project"


## Bundled Resources

### references/

- `crucible-structure.md` — Complete 36-beat structure
- `question-sequences.md` — Full question bank by document
- `forge-point-rules.md` — Strand convergence mechanics
- `dark-mirror-guide.md` — Antagonist design
- `mercy-engine-guide.md` — Mercy/payoff mechanics

### assets/templates/

- Document templates for generation

### scripts/

- `init_project.py` — Initialize project
- `save_state.py` — Save progress
- `load_state.py` — Load progress
- `compile_documents.py` — Generate all documents

平台分布

Codex

35.36%
按下载量换算28

Claude

33.77%
按下载量换算26

Cursor

18.89%
按下载量换算15

Gemini CLI

8.76%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills