复制安装命令
用 Codex 或 Claude 安装复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它先审查 Skill 页面再帮你安装。
复制前请先查看来源、License 和安全提示。
Turn a novel in any language into a source-grounded, fully playable game.
用 Codex 或 Claude 安装复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它先审查 Skill 页面再帮你安装。
复制前请先查看来源、License 和安全提示。
来源文件:README.md
Turn a novel in any language into a source-grounded, fully playable game.
NovelToGame is an open-source Agent Skills toolkit for Claude Code, Codex, and Kimi Code. It turns novel adaptation into a staged workflow: source analysis, concept selection, world and art direction, implementation, and runtime QA.
Bring a novel in any language and choose the target runtime. Generated artifacts follow the requested language; build and QA stay on the chosen platform instead of silently falling back to an easier substitute.
中文 · Play Online · Quick Start · Workflow · Skills · Artifacts · Contributing
Three playable adaptations, each available in a browser right now and linked to the case study behind it: source provenance, concept trade-offs, game and art direction, runnable source, and evidence from the playable paths.
One wave of a fan blew you fifty thousand li. Take the mountain back one turn at a time.
Command Wukong's party through the three borrowings of the Banana-Leaf Fan: read the five-element wheel, follow a fire-vein treasure map, decide when to press deeper or bank the haul, transform your way in where force will not work, and turn a demon king who outclasses you into a rainstorm over the Mountain of Flames.
Play in browser · Read the case study · design estimate: 45–90 min · all ages · playable prototype
Choose whose door you enter tonight. Find out whose door knocks in the morning.
Twenty days, five courtyards. Keep silver, influence, reputation, exposure, and household strain in balance; respect each woman's terms; build trust through shared crises; and face a final ledger shaped by what everyone chose and remembers.
Play in browser · Read the case study · design estimate: 60–90 min · 18+ · playable prototype
A real-time first-person 3D field-photography game adapted from Arthur Conan Doyle's The Lost World. Cross a connected plateau, observe a living Iguanodon family, expose four glass plates under aerial pressure, and return with the views that survived.
Play the full expedition on desktop, or watch the 15-second gameplay preview on other devices.
https://github.com/user-attachments/assets/27819247-4e4d-4bf0-8f0f-43d4125c4d45
Play in your browser — no install · Read the case study · Share feedback · 1–3 min run · desktop WebGL2 · playable prototype
A one-line “turn this book into a game” prompt often produces a generic reskin or a clickable plot summary. NovelToGame keeps the adaptation traceable and gives each major decision a clear owner:
| Agent CLI | Install | Invoke |
|---|---|---|
| Claude Code | npx skills add zenstory-ai/novel-to-game -g -y -a claude-code -s '*' | /novel-to-game |
| Codex | npx skills add zenstory-ai/novel-to-game -g -y -a codex -s '*' | $novel-to-game |
| Kimi Code | npx skills add zenstory-ai/novel-to-game -g -y -a kimi-code-cli -s '*' | /skill:novel-to-game |
Install adapters for all three CLIs on the same machine:
npx skills add zenstory-ai/novel-to-game -g -y -s '*' \
-a claude-code -a codex -a kimi-code-cli
Cloning the repository also enables project-local skill discovery in all three CLIs.
Give the agent a novel file, directory, or link:
Use novel-to-game quick to adapt this novel into a fully playable game.
Recommend the target platform, genre, and engine from the source, and keep the first build to about 15 minutes.
Let the player enter the world as an original character with a new playable route through its conflict.
When you want an interactive story rather than a systems game, say so. That locks the narrative-led experience profile, so concept, design, and QA judge continuous scenes, character dialogue, testimony, and key choices instead of applying rounds, cards, and resource bars:
Use novel-to-game quick to adapt this novel into an interactive story.
Carry the experience with continuous scenes, character dialogue, testimony, and key choices.
Keep variables as hidden causal tags rather than a visible stat panel.
Key choices must change later scenes, character attitudes, and the ending, and be named back in later text.
The narrative track lowers no standard: it still needs a recognizable gameplay precedent, the same three-phase arc, and the same hard vetoes. Only the expression changes -- new people to question, new ways to press a contradiction, and attitudes that shifted because of what you did earlier.
quick is the low-friction option: the agent drafts sensible defaults, asks only about materially branching or safety-sensitive choices, compares three concepts, and continues through design, build, and QA. Every project runs one minimum QA path covering real startup, rendering, input, a complete loop, an outcome, restart, and explicit limitations. It does not require a human playtest or a separate approval report. Choose director when you want to pick the concept yourself.
/plugin marketplace add zenstory-ai/novel-to-game
/plugin install novel-to-game@novel-to-game-skills
/novel-to-game:novel-to-game quick
codex plugin marketplace add zenstory-ai/novel-to-game
codex plugin add novel-to-game@novel-to-game-skills
/plugins install https://github.com/zenstory-ai/novel-to-game
/reload
/skill:novel-to-game quick
The orchestrator locks PRODUCT_BRIEF.md, then hands the adaptation through separately owned decisions. Concept, experience/level design, and art direction remain distinct. After world design, a risk-matched whitebox tests the hardest causal, systemic, spatial, or control question before art production; its findings return to the design owner rather than becoming a new QA gate.
Novel → Source analysis → Concept → World design → Risk-matched whitebox ↺ → Art direction → Production build → QA → Playable game
The whitebox runs only the narrow model/replay check needed for its declared risk. The production build targets the chosen runtime and prepares one authoritative verification entry point; QA runs it once and records the six minimum player-visible effects with real execution evidence. Capability-specific regression checks run only when that capability is actually adopted. No human-playtest gate or duplicate QA report is required. Source identity, public hosting, marketing, rights, subjective fun, and publication quality are not machine-proven by this QA record.
| Skill | Responsibility |
|---|---|
novel-to-game | Confirm requirements, choose a mode, orchestrate stage handoffs, and recover progress |
novel-game-analyze | Extract cited rules, verbs, spaces, agents, systems, and signature moments |
game-concept | Generate three materially different directions, reject invalid options, and choose one |
game-world-design | Define the player promise, core loop, world response, systems, levels, failure, and outcomes |
game-art-direction | Define camera, composition, visual grammar, colour, light, materials, HUD, motion, and sound |
game-build | Build a risk-matched whitebox, then implement the approved production candidate without redesigning it |
game-qa | Verify commands, states, screenshots, and real play paths without overstating subjective results |
Each run creates a compact, self-contained adaptation workspace:
game-adaptations/<project>/
PRODUCT_BRIEF.md
analysis/SOURCE_BIBLE.md
concepts/CONCEPT.md
design/GAME_DESIGN.md
design/ART_DIRECTION.md
build/BUILD_BRIEF.md
build/app/
qa/verification.json
_progress.md
Core design documents remain independent of any single model or game engine. The approved target runtime determines the implementation and QA environment.
Reproducible bugs, skill gaps backed by evidence, and example proposals that demonstrate a distinct adaptation lesson are welcome. Read the contribution guide and use the repository's structured issue and pull request templates.
NovelToGame is released under the MIT License.
Thanks to the linux.do community for early feedback and support.
name: novel-game-analyze
description: "Deconstruct a novel for game adaptation. Compress a raw novel, deconstruction library, or writing project into a SOURCE_BIBLE with cited textual evidence, extracting world rules, player verbs, spaces, character will, systems, and visual anchors — without inventing a genre yet. Use for gameable book analysis, analyze a novel for game adaptation, extract game systems from a book. 游戏化拆书。把原始小说、拆文库或写作工程压缩成有必要原文证据的 SOURCE_BIBLE,提取世界规则、玩家动词、空间、角色意志、系统和视觉锚点,不提前发明游戏类型。用于游戏化拆书、分析小说的游戏改编基础等需求。"建立游戏设计所需的事实层,不评价文笔,也不在此阶段选择游戏类型。
产物语言由 PRODUCT_BRIEF.md 锁定;未锁定时跟随对话语言,不默认产出中文。
原文可以是任意语言。引用保留原文;策划语言或界面语言与原文不同时附必要译文,并建立角色名、 地名、物件和规则的统一术语表;礼制、宗教和叙事惯例不替换成另一文化的近似类型标签。
优先复用现有写作工程或全量拆文产物,只回原文补关键证据;只有原始小说时才执行 本 skill 的全书管道。已有分析若只覆盖黄金三章或快速预览,只能帮助判断开篇,不能 证明理解了全书。除非用户明确只改编某一卷或片段,否则来源未覆盖全文时不得进入概念设计。
不因目录看起来完整就直接信任;按协议「来源清点」清点来源与已有拆文后,再判定是否算全量。
若存在 PRODUCT_BRIEF.md,先读取:分级、核心幻想与已锁类型决定取证优先级——优先为
核心幻想涉及的动作、关系、场面取证;对已锁类型需要而原作未系统提供的体系(如属性相克、
数值成长),在设计空白中写明「原作无系统化的对应体系,属下游发明」,供下游改编边界如实
引用。仍不在本阶段发明或评价类型。
建立边界 → 逐单元提取 → 覆盖门禁 → 分层聚合 → 回查原文 → 压缩交付。各步做法以 gameability-protocol.md 为准,此处只列顺序,不重述。选定改编段落的核对在概念选定后二次 进入时做,产出逐事实表(见「输出」)。工作卡只服务覆盖、恢复和复核,不把过程日志包装成策划成果。
可以在运行时并行处理互不重叠的章节批次,但每批必须使用同一提取契约,聚合和事实回查 保持串行。环境不支持并行时逐批串行完成,不能改成抽样。
只保留会影响玩家动作、世界响应或视听身份的内容;字段以协议「逐章游戏化卡」的十行为准,不得 只取其中一部分,该表即并行批次共用的提取契约。另记不可改的原作事实、授权风险和原作未定义的设计空间。
若作品反复出现命令、委托或代办,同样落回既有十行:记录谁有资格下令、谁实际执行、所需权限与物件、 可能拒绝或走样的环节、谁会见证,以及什么债务或后果会延期返回。此阶段只提取身份动作的原作依据, 不预设运行时必须有自由输入,也不把一句命令直接等同于事情已经发生。
为硬规则、关键角色目标、玩家身份依据、转折结局和标志性锚点记录章节或文件位置。 普通候选项不必机械附带置信度对象;推断必须明确标成推断或开放设计空间。
生成一份 analysis/SOURCE_BIBLE.md,包含来源范围、全书覆盖、世界身份、原作事实与
证据、改编边界、玩家身份候选(PRODUCT_BRIEF 已锁身份时改为该身份的动作空间验证)、
可玩动词、可下达/承接的身份动作及其执行阻力、世界模型、空间与角色能动性、文化语境、
术语表(语言条件见上)、视觉锚点、风险和设计空白。全书覆盖按连续范围列出每一阶段的关键变化、
强弱节奏、反例/回收与游戏化意义,
表头写明来源单元数与成功数,使覆盖集合与来源单元集合完全一致。短表格优先于重复文档、
结构化数据和证据索引。
改编边界必须是表格:条目 | 边界类型 | 证据位置,边界类型只用协议的四种标签;涉及
名分、座次、官阶或队伍序列的作品必须附人物名册行(姓名 | 身份/次序 | 结局 | 证据回目),
作为下游各交接门逐项比对的名册基准。
概念阶段选定改编段落后,二次进入本 skill 为该段落补一张逐事实表并追加进 SOURCE_BIBLE:
事实 | 证据位置 | 边界类型,含入镜角色的外观要点(服饰、兵器、体量及位置;原文无描写
写「原文未明确」)。此表是下游交接门的段落级比对基准;下游仍可回原文补证,发现出入时
回上游修正此表,不得各自改写。追加这类原作事实与证据是概念选定后对 SOURCE_BIBLE 的
唯一允许更新——SOURCE_BIBLE 不得包含下游发明物,游戏状态键名、回合或日程编号、
scene_id、Hub/界面结构、QA 与范围计划只属于 GAME_DESIGN,不得把既有事实改写成
切片说辞。
来源范围只记录会影响复核的版本、覆盖范围和授权边界。下载日期、字节数、校验值和 本地路径只有在复现契约确实需要时才写,不把获取日志当策划内容。
完成前确认:覆盖数与来源数一致且无未处理失败;核心规则可追溯;改编边界成表且证据位置 齐全;能回答玩家可能做什么、在哪里做、世界如何回应;理解开篇、中段、结局和跨段回收; 没有把一次性剧情冒充核心循环,也没有提前选定游戏类型。
评论 (0)
暂无评论,成为第一个评论者吧!