@oh-story/dsh 0.1.15 → 0.1.17

package/lib/novel-to-game/README.md

@@ -31,7 +31,7 @@
   <a href="https://github.com/zenstory-ai/novel-to-game/issues"><img alt="GitHub Issues" src="https://img.shields.io/badge/GitHub%20Issues-181717?style=for-the-badge&logo=github&logoColor=white"></a>
 </p>
 
-NovelToGame is seven open-source skills for the coding agent you already use (Claude Code, Codex, or Kimi Code). Give it a novel and, if you have one in mind, a target platform or engine. It reads the whole book, chooses a game direction it can defend, designs the world and the look, builds for the runtime the brief locks, and proves the build plays in one recorded run. You get a playable build at the finish level you asked for, plus the design documents behind it. The three public examples below play in a browser right now, no install.
+NovelToGame is seven open-source skills for the coding agent you already use (Claude Code, Codex, or Kimi Code). Give it a novel and, if you have one in mind, a target platform or engine. It reads the whole book, chooses a game direction it can defend, designs the world and the look, builds for the runtime the brief locks, and proves the build plays in one recorded run. You get a playable build at the finish level you asked for, plus the design documents behind it. The three examples below are browser games and include links to their playable builds.
 
 ## Play Online
 
@@ -75,13 +75,13 @@ https://github.com/user-attachments/assets/5ad62a58-1abe-4d73-b86e-1c8c6563fdb2
 
 ## What it is
 
-There is no GPU, no hosted service, and no bundled engine: the skills run on the model of whichever agent you installed them into. A one-line "turn this book into a game" prompt usually yields a reskin or a clickable plot summary, so the work is split into stages. Each stage owns one document, and every decision in it has to cite evidence:
+The skills run on your coding agent's model and organize adaptation into stages. Each stage owns a document, with source evidence separated from design choices:
 
-- **Every design decision cites the novel.** Analysis reads the whole book and writes a `SOURCE_BIBLE` where every hard rule, key character goal, turning point and ending, and signature anchor carries its chapter or file location, and every adaptation-boundary row cites its evidence. Source facts are labelled `immutable`, `adaptable`, `open`, or `conflicted`; anything the novel does not define is marked a design invention, never smuggled in as fact.
+- **Key source facts carry citations.** Analysis reads the whole book and writes a `SOURCE_BIBLE` where every hard rule, key character goal, turning point and ending, and signature anchor carries its chapter or file location, and every adaptation-boundary row cites its evidence. Source facts are labelled `immutable`, `adaptable`, `open`, or `conflicted`; anything the novel does not define is marked a design invention, never smuggled in as fact.
 - **The game shape is a documented decision, not a default.** The concept stage compares real alternatives and kills directions with named hard vetoes (the core loop ignores the novel's central tension; strip the proper nouns and a generic template remains; the player only spends resources to release a fixed plot). Concept, world design, and art direction are written by different stages into different documents, and the build may not quietly redesign them.
 - **A borrowed genre comes with its shell.** Players recognise a genre by more than its core mechanic, so a systems game lists which parts of the genre's shell (entering the world, objective guidance, progression and reasons to return) it implements, degrades, or drops and why, and what the first minute shows. A Fantasy Westward Journey-style RPG is not shipped as a bare chain of fights.
 - **Built for the runtime the brief locks.** The product brief, the one-page requirements sheet the agent drafts first (in `quick` it stops for your decision only on blocking items such as platform, rights, or adult content), locks platform, engine, target runtime, and the runtime actually available for testing. A missing toolchain may not silently become a web build. The design documents are engine-independent Markdown you own.
-- **QA is one real run, six checks, and honest limitations.** The build is launched on the tested runtime and one complete execution proves `launch`, `render`, `input`, `coreLoop`, `outcome`, and `restart`. Fun, balance, other browsers, and rights are written as limitations, never as PASS.
+- **QA is one real run, six checks, and honest limitations.** The build is launched on the tested runtime and one complete execution proves `launch`, `render`, `input`, `coreLoop`, `outcome`, and `restart`. Untested environments are listed as limitations; this PASS does not establish fun, balance or rights clearance.
 - **Voice is opt-in and generated at build time by default.** No line goes to a text-to-speech provider unless art direction chooses voice; runtime synthesis needs the product brief's explicit approval, and the novel and design documents are never uploaded.
 
 Interactive fiction is a first-class track: continuous scenes, dialogue, testimony, and key choices can carry the whole game, held to the same agency and evidence requirements as a systems game. On a character route, a key choice targets that character's own wound or agenda, and the relationship shows in reactions, forms of address, the places that open to you and what you are trusted with, not in an on-screen number.
@@ -90,8 +90,7 @@ Interactive fiction is a first-class track: continuous scenes, dialogue, testimo
 
 ### ClawHub
 
-ClawHub distribution is governed by the repository's [explicit publish inventory](.clawhub/publish.json), with discovery through [ClawHub](https://clawhub.ai/). A skill-specific link is added here only after its publisher, version, and anonymous accessibility are verified, so this README never presents a nonexistent listing as live.
-
+ClawHub distribution is governed by the repository's [explicit publish inventory](.clawhub/publish.json), with discovery through [ClawHub](https://clawhub.ai/).
 
 Prerequisite: you already use Claude Code, Codex, or Kimi Code, and `npx` (Node.js) runs in your terminal.
 
@@ -140,7 +139,7 @@ codex plugin add novel-to-game@novel-to-game-skills
 
 </details>
 
-> Changes are in [CHANGELOG.md](CHANGELOG.md) and [Releases](https://github.com/zenstory-ai/novel-to-game/releases). The repository moved from `worldwonderer/novel-to-game` to `zenstory-ai/novel-to-game`; install and marketplace commands from older docs no longer resolve, and the commands above are current.
+> Changes are in [CHANGELOG.md](CHANGELOG.md) and [Releases](https://github.com/zenstory-ai/novel-to-game/releases).
 
 ## See what it produces
 
@@ -360,7 +359,7 @@ Keep the choices and outcomes bounded; show each option's evidence, cost, visibl
 Label allowed additions and unresolved questions. Deliver design notes only—do not build, run QA, or claim a finished runtime.
 ```
 
-`quick` is the default: the agent drafts a product brief with sensible defaults, asks only about choices that materially change direction or touch safety, compares meaningful alternatives, and continues through design, build, and QA. Choose `director` when you want to pick the concept yourself; the agent stops at the concept stage with candidates and a recommendation, unless you have already named a direction.
+`quick` is the default: the agent drafts a product brief with sensible defaults, asks only about choices that materially change direction or touch safety, compares meaningful alternatives, and continues through design, build, and QA. Choose `director` when you want to make the key calls yourself: the agent stops with candidates and a recommendation at three points (the concept, what to cut after the whitebox playtest, and the visual target images) unless you have already decided.
 
 ## Workflow
 
@@ -368,7 +367,7 @@ Label allowed additions and unresolved questions. Deliver design notes only—do
 Novel → Source analysis → Concept → World design → Risk-matched whitebox ↺ → Art direction → Production build → QA → Playable game
 ```
 
-After world design, a whitebox (a rough build without final art) tests only the single largest design risk before full art production: narrative causality can be checked in text, but real-time control, space, physics, or camera cannot, and what it shows goes back to the design stage. The production build targets the approved runtime and prepares one authoritative verification command. QA may diagnose, fix, and rerun, but the final record binds all six checks to the same complete run; findings flow back to the product, design, art, or build document that owns them.
+After world design, a whitebox (a rough build without final art) tests only the single largest design risk before full art production: narrative causality can be checked in text, but real-time control, space, physics, or camera cannot, and what it shows goes back to the design stage. The whitebox playtest also shows which part players keep coming back to, so the story, dialogue and level budget is revised against it, and the brief's production budget is checked against the remaining scope there and again halfway through production. The production build targets the approved runtime and prepares one authoritative verification command. QA may diagnose, fix, and rerun, but the final record binds all six checks to the same complete run; findings flow back to the product, design, art, or build document that owns them.
 
 ## Skills
 
@@ -409,7 +408,7 @@ Say so in the request, as the second example above does. That locks `experienceP
 
 ### How long does a full run take, and what does it cost?
 
-Longer than one prompt. The agent reads the whole novel, writes five design documents, builds, and runs QA, and each stage is a separate skill call, so a full run is long and may span more than one session. Token spend is whatever your coding agent bills; nothing else is paid unless art direction opts into image or voice generation (see the GPU question below). The public examples' progress files do not record wall-clock time or token counts, so this README does not quote a number.
+The agent reads the whole novel, writes the adaptation documents, builds, and runs QA through separate skill calls. A full run may span more than one session. Token usage is billed by your coding agent. The product brief records the budget you give it (model quota, time or money), and the run checks the remaining scope against it after the whitebox and halfway through production, cutting content before the budget runs out. Generation services, engine licences or hosting may have separate costs; the requirements stage identifies paid services before use. The public examples' progress files do not record wall-clock time or token counts, so this README does not quote a number.
 
 ### The session ended halfway. Can it continue?
 
@@ -421,7 +420,7 @@ You choose, and the product brief locks it: platform, production engine, target
 
 ### Do I need a GPU or my own model? Who pays for images and voice?
 
-No GPU and no separate model: analysis, design, and code are written by the model of the coding agent you already use. External services are opt-in. Generated images are used only when art direction selects them, with the tool chosen in your environment after checking capability, licence, and cost (the Jin Ping Mei example records most of its images, generated with Codex's built-in image tool, batch by batch in `build/art/generated-art.json`). Voice is off unless art direction chooses it; the default is build-time generation into local assets, only the per-line dialogue and any pronunciation notes are sent, and the novel and design documents are never uploaded. Paid services are one of the things the agent stops to ask about before it starts. See the [TTS production contract](skills/game-build/references/tts-production-contract.md) (Chinese).
+For analysis, design and coding, the skills use your coding agent's model. If that model is hosted, these steps require no extra local model or GPU. Running the resulting game follows its target runtime's hardware requirements. External services are opt-in. Generated images are used only when art direction selects them, with the tool chosen in your environment after checking capability, licence, and cost (the Jin Ping Mei example records most of its images, generated with Codex's built-in image tool, batch by batch in `build/art/generated-art.json`). Voice is off unless art direction chooses it; the default is build-time generation into local assets, only the per-line dialogue and any pronunciation notes are sent, and the novel and design documents are never uploaded. Paid services are one of the things the agent stops to ask about before it starts. See the [TTS production contract](skills/game-build/references/tts-production-contract.md) (Chinese).
 
 ### How do I know the game it built actually runs? What does the QA PASS mean?
 

package/lib/novel-to-game/README_ZH.md

@@ -1,4 +1,4 @@
-<!-- Last synced with README.md: 2026-10-02 -->
+<!-- Last synced with README.md: 2026-10-19 -->
 
 <p align="center">
   <img src="https://zenstory.ai/brand/zenstory-ai-mark.svg" alt="" width="76" height="76">
@@ -33,11 +33,11 @@
   <a href="https://github.com/zenstory-ai/novel-to-game/issues"><img alt="GitHub Issues" src="https://img.shields.io/badge/GitHub%20Issues-181717?style=for-the-badge&logo=github&logoColor=white"></a>
 </p>
 
-NovelToGame 是装进你已在用的编码 Agent（Claude Code、Codex、Kimi Code）的七个开源 skills。把一本小说交给它——如果已有想法，再加上目标平台或引擎。它会读完整本书，选一个站得住的游戏方向，设计世界和画面，在简报锁定的运行环境里构建，并用一次有记录的运行证明这个版本能玩。你拿到的是一款达到你所要求成色的可玩构建，加上它背后的设计文档。下面三个公开示例现在就能在浏览器里打开，无需安装。
+NovelToGame 是装进你已在用的编码 Agent（Claude Code、Codex、Kimi Code）的七个开源 skills。把一本小说交给它——如果已有想法，再加上目标平台或引擎。它会读完整本书，选一个依据充分的游戏方向，设计世界和画面，在简报锁定的运行环境里构建，并用一次有记录的运行证明这个版本能玩。你拿到的是一款达到你所要求成色的可玩构建，加上它背后的设计文档。下面三个示例都是浏览器游戏，并提供可玩版本的链接。
 
 ## 在线试玩
 
-每款游戏都链到它背后的改编工作区：原文来源、概念与被否决的方向、游戏与美术方向、可运行源码，以及来自可玩路径的 QA 记录。
+每款游戏都链接到它背后的改编工作区：原文来源、概念与被否决的方向、游戏与美术方向、可运行源码，以及来自可玩路径的 QA 记录。
 
 ### 西游记 · 三借芭蕉扇
 
@@ -77,23 +77,22 @@ https://github.com/user-attachments/assets/5ad62a58-1abe-4d73-b86e-1c8c6563fdb2
 
 ## 这是什么
 
-不需要 GPU，没有托管服务，也不内置游戏引擎：这套 skills 用的就是你把它装进哪个 Agent、那个 Agent 的模型。只给模型一句“把这本书做成游戏”，得到的通常是换皮或可点击的剧情摘要，所以工作被拆成几个阶段：一个阶段负责一份文档，每个决定都要有证据：
+这套 skills 由你已在用的编码 Agent 的模型执行，按阶段组织改编工作。每个阶段负责一份文档，明确区分原著证据与设计选择：
 
-- **每个设计决定都引用原著。** 拆解阶段读完全书，写出 `SOURCE_BIBLE`：硬规则、关键角色目标、转折结局和标志性锚点都标注章回或文件位置，改编边界表每行带证据位置。原作事实分为 `immutable`、`adaptable`、`open`、`conflicted` 四种边界；原著没有定义的部分标为设计发明，不会被悄悄写成事实。
+- **关键原作事实都有引证。** 拆解阶段读完全书，写出 `SOURCE_BIBLE`：硬规则、关键角色目标、转折结局和标志性锚点都标注章回或文件位置，改编边界表每行带证据位置。原作事实分为 `immutable`、`adaptable`、`open`、`conflicted` 四种边界；原著没有定义的部分标为设计发明，不会被悄悄写成事实。
 - **游戏形态是一个有记录的决定，不是默认值。** 概念阶段比较真实的替代方向，用具名的硬否决淘汰不合格方向（核心循环与原作中心张力无关；删掉专有名词后只剩通用模板；玩家只是花资源放行既定剧情）。概念、世界设计与美术方向分别由不同阶段写进各自的文档，构建阶段不得静默重做。
 - **借成熟品类，就连同它的外壳一起借。** 玩家认出一类游戏，靠的不只是核心玩法；系统玩法项目要用一张短表写清该品类的外壳（进入世界、目标指引、成长与回访理由）哪些实现、哪些降级、哪些不做以及为什么，并写清第一分钟玩家看到什么。《梦幻西游》式的 RPG 不会只交出一串战斗关卡。
 - **面向简报锁定的运行环境构建。** 产品简报（Agent 最先起草的一页需求单；`quick` 模式下只在平台、权利、尺度等阻断项上停下等你裁决）锁定平台、引擎、目标运行时和实际可用的测试运行时。工具链缺失时不得悄悄改做网页。设计文档是不绑定引擎的 Markdown，归你所有。
-- **QA 是一次真实运行、六项检查和如实写下的限制。** 构建在测试运行时里真正启动，一次完整运行证明 `launch`、`render`、`input`、`coreLoop`、`outcome`、`restart` 六项。趣味、平衡、其他浏览器和权利只写成限制项（limitation），不写成 PASS。
+- **QA 是一次真实运行、六项检查和如实写下的限制。** 构建在测试运行时里真正启动，一次完整运行证明 `launch`、`render`、`input`、`coreLoop`、`outcome`、`restart` 六项。未测试的环境写入限制项（limitation）；这个 PASS 不证明趣味、平衡或已经清权。
 - **语音默认关闭，默认在构建期生成。** 除非美术方向选用语音，没有任何台词会发给语音合成服务；运行时合成须经产品简报明确批准；小说和设计文档从不上传。
 
-互动叙事是一等赛道：连续场景、对白、证词与关键选择可以承载整个游戏，但要满足与系统玩法同样的能动性和证据要求。人物路线的关键选择针对该人物自己的心结或议程，关系进展由人物反应、称呼、可进入的场所与可托付的事呈现，而不是屏幕上的数值。
+这套 skills 完整支持互动叙事：连续场景、对白、证词与关键选择可以承载整个游戏，但要满足与系统玩法同样的能动性和证据要求。人物路线的关键选择针对该人物自己的心结或议程，关系进展由人物反应、称呼、可进入的场所与可托付的事呈现，而不是屏幕上的数值。
 
 ## 安装
 
 ### ClawHub
 
-ClawHub 分发由仓库中的 [显式发布清单](.clawhub/publish.json) 管理，并通过 [ClawHub](https://clawhub.ai/) 提供发现入口。具体 skill 链接只有在发布者、版本和匿名访问均验证后才会写入本文档，避免把尚不存在的条目当成已上架。
-
+ClawHub 分发由仓库中的 [显式发布清单](.clawhub/publish.json) 管理，并通过 [ClawHub](https://clawhub.ai/) 提供发现入口。
 
 前提：你已经在用 Claude Code、Codex 或 Kimi Code，终端能运行 `npx`（Node.js）。
 
@@ -142,7 +141,7 @@ codex plugin add novel-to-game@novel-to-game-skills
 
 </details>
 
-> 变更见 [CHANGELOG.md](CHANGELOG.md) 与 [Releases](https://github.com/zenstory-ai/novel-to-game/releases)。仓库已从 `worldwonderer/novel-to-game` 迁至 `zenstory-ai/novel-to-game`，旧文档里的安装与 marketplace 命令已失效，以上面的命令为准。
+> 变更见 [CHANGELOG.md](CHANGELOG.md) 与 [Releases](https://github.com/zenstory-ai/novel-to-game/releases)。
 
 ## 看看它的输出
 
@@ -361,7 +360,7 @@ PASS 只在记录下来的本地桌面浏览器里证明这六项效果。
 标明允许新增的设定与待决问题。只交付设计说明，不构建、不运行 QA，也不声称运行成品已经完成。
 ```
 
-`quick` 是默认模式：Agent 先用合理默认值起草产品简报，只追问会改变方向或涉及安全的选择，比较有效替代后继续设计、构建和 QA。想自己选概念时用 `director`：Agent 会在概念阶段停下，给出候选和推荐；你已明确选定方向时不再停靠。
+`quick` 是默认模式：Agent 先用合理默认值起草产品简报，只追问会改变方向或涉及安全的选择，比较有效替代后继续设计、构建和 QA。想自己做关键决定时用 `director`：Agent 会在概念方向、白盒试玩后的内容取舍和视觉目标图三处停下，给出候选和推荐；你已经做出决定的地方不再停靠。
 
 ## 工作流
 
@@ -369,7 +368,7 @@ PASS 只在记录下来的本地桌面浏览器里证明这六项效果。
 小说 → 游戏化拆解 → 游戏概念 → 世界设计 → 风险白盒 ↺ → 美术方向 → 正式构建 → QA → 可玩游戏
 ```
 
-世界设计完成后，先用白盒（不要求最终美术的粗略构建）只验证最大的一个设计风险，再进入完整美术生产：叙事因果可以用文本验证，实时操作、空间、物理或镜头不能；观察结果交回设计阶段。正式构建面向批准的运行环境，并准备一条权威验证命令。QA 可以诊断、修复和复跑，但最终记录把六项检查绑定到同一次完整运行；发现的问题按产品、设计、美术、构建各自的文档归属回流。
+世界设计完成后，先用白盒（不要求最终美术的粗略构建）只验证最大的一个设计风险，再进入完整美术生产：叙事因果可以用文本验证，实时操作、空间、物理或镜头不能；观察结果交回设计阶段。白盒试玩还要看出玩家反复回来做的是哪一部分，剧情、对白和关卡的内容预算据此修订；简报里的制作预算也在这一步、以及正式生产进行到一半时，与剩余范围各对照一次。正式构建面向批准的运行环境，并准备一条权威验证命令。QA 可以诊断、修复和复跑，但最终记录把六项检查绑定到同一次完整运行；发现的问题按产品、设计、美术、构建各自的文档归属回流。
 
 ## Skills
 
@@ -410,7 +409,7 @@ game-adaptations/<project>/
 
 ### 完整跑一次要多久、花多少钱？
 
-不止一句提示词的事。Agent 要读完整本小说、写五份设计文档、构建并跑 QA，每个阶段都是一次独立的 skill 调用，所以一次完整运行很长，可能跨多次会话。token 费用按你所用编码 Agent 的计费；除非美术方向选用图片或语音生成（见下面 GPU 一条），不产生其他费用。公开示例的进度文件没有记录耗时和 token 数，所以这里不给具体数字。
+Agent 要通过不同的 skill 调用读完整本小说、写改编文档、构建并跑 QA，完整运行可能跨多次会话。token 用量按你所用编码 Agent 计费。产品简报会记下你给出的预算（模型额度、工时或费用），流程在白盒之后和正式生产进行到一半时各用它对照一次剩余范围，在预算耗尽前先删减内容。生成服务、引擎许可或托管可能另有费用，需求确认阶段会在使用付费服务前说明。公开示例的进度文件没有记录耗时和 token 数，所以这里不给具体数字。
 
 ### 会话中途断了，能接着做吗？
 
@@ -418,11 +417,11 @@ game-adaptations/<project>/
 
 ### 它面向哪个引擎或平台构建？我的工具链缺失时会不会退回网页？
 
-由你决定，并由产品简报锁定：平台、生产引擎、目标运行时，以及实际可用的测试运行时。目标工具链不可用时，构建不得自动改做网页；只能使用简报里已批准的替代运行时，并分开记录 `targetRuntime`、`testedRuntime` 和未覆盖项。QA 把替代运行视为从不证明目标平台已通过。三个公开示例都是浏览器游戏（两个零依赖静态应用和一个 Three.js + Vite 构建），是因为它们的简报锁定了这样，不是因为这是默认值。目前没有任何公开示例面向原生引擎或主机；Godot、Unity 或移动端构建取决于你环境里的工具链，并按同样方式记录，未测到的部分写成限制项。见 [`game-build`](skills/game-build/SKILL.md) 与 [QA 契约](skills/game-qa/references/qa-contract.md)。
+由你决定，并由产品简报锁定：平台、生产引擎、目标运行时，以及实际可用的测试运行时。目标工具链不可用时，构建不得自动改做网页；只能使用简报里已批准的替代运行时，并分开记录 `targetRuntime`、`testedRuntime` 和未覆盖项。替代运行的 QA 结果只覆盖实际测试环境。三个公开示例都是浏览器游戏（两个零依赖静态应用和一个 Three.js + Vite 构建），它们的简报选择了浏览器作为目标运行环境。目前没有任何公开示例面向原生引擎或主机；Godot、Unity 或移动端构建取决于你环境里的工具链，并按同样方式记录，未测到的部分写成限制项。见 [`game-build`](skills/game-build/SKILL.md) 与 [QA 契约](skills/game-qa/references/qa-contract.md)。
 
 ### 需要 GPU 或自己的模型吗？图片和语音的钱谁出？
 
-不需要 GPU，也不需要单独的模型：拆解、设计和代码都由你已在用的编码 Agent 的模型完成。外部服务是可选项。生成图片只在美术方向选择它时使用，工具在你的环境里按能力、许可与费用核实后选择（金瓶梅示例的图由 Codex 内置的图像工具生成，大部分在 `build/art/generated-art.json` 里按批次记录）。语音默认关闭，除非美术方向选用；默认在构建期生成为本地资产，只发送逐句台词与必要读音，从不上传小说和设计文档。付费服务是 Agent 开始前会停下来问你的问题之一。见[语音资产生产合同](skills/game-build/references/tts-production-contract.md)。
+拆解、设计和代码都由你已在用的编码 Agent 的模型完成。使用托管模型时，这些步骤不需要额外的本地模型或 GPU；运行生成的游戏仍需满足目标运行环境的硬件要求。外部服务是可选项。生成图片只在美术方向选择它时使用，工具在你的环境里按能力、许可与费用核实后选择（金瓶梅示例的图由 Codex 内置的图像工具生成，大部分在 `build/art/generated-art.json` 里按批次记录）。语音默认关闭，除非美术方向选用；默认在构建期生成为本地资产，只发送逐句台词与必要读音，从不上传小说和设计文档。付费服务是 Agent 开始前会停下来问你的问题之一。见[语音资产生产合同](skills/game-build/references/tts-production-contract.md)。
 
 ### 怎么知道它做出来的游戏真的能跑？QA 里的 PASS 是什么意思？
 

package/lib/novel-to-game/VERSION

@@ -1 +1 @@
-0.5.0
+0.6.0

package/lib/novel-to-game/manifest.json

@@ -2,10 +2,10 @@
   "schemaVersion": 1,
   "upstream": {
     "repository": "https://github.com/zenstory-ai/novel-to-game.git",
-    "commit": "e2695b720591335321738039b869f418593b1985",
-    "releaseVersion": "0.5.0"
+    "commit": "7758caa8d9578f19780349e7b4e4115ef831fd7a",
+    "releaseVersion": "0.6.0"
   },
-  "generatedAt": "2026-10-03T14:32:39.565Z",
+  "generatedAt": "2026-10-19T08:12:54.284Z",
   "skills": [
     "game-art-direction",
     "game-build",
@@ -26,17 +26,17 @@
     },
     {
       "path": "README.md",
-      "sha256": "b8fa61a8ad38c58d8074aa40f3039107cfd13537cd551a535fc27e328da5b1ef",
-      "bytes": 38659
+      "sha256": "dc046f8d7f8cee9af8cc3882b51b274ba04b51d04ff3b3eafaf555b1cc10f1e0",
+      "bytes": 38777
     },
     {
       "path": "README_ZH.md",
-      "sha256": "068853d6beb986a8ae18a9bbfa91e433cdac16fd35d25e4dfab1db3e66e0dc89",
-      "bytes": 34757
+      "sha256": "5332bdb07c7744e8b556d253d28d1c3edd8e7feb04990d8a109fe19a643ff6ca",
+      "bytes": 34810
     },
     {
       "path": "VERSION",
-      "sha256": "a0ad50640139a524f86591157145bf1d4674ea9c065bfea21558f0f7b04430c8",
+      "sha256": "97a5d6e540a6120d09c39ddb667e0e8e2bac7fcc2ffef813b82e08fd60365bce",
       "bytes": 6
     },
     {
@@ -576,8 +576,8 @@
     },
     {
       "path": "skills/game-art-direction/SKILL.md",
-      "sha256": "79aada543bbc876e7d79e4989cc87067ed7e01d613933157689855c7b8d14b0b",
-      "bytes": 3227
+      "sha256": "73f95a5e747709a4f35b2703fabc2f00f48fb63b93a1aff002251e9af06dff6a",
+      "bytes": 3245
     },
     {
       "path": "skills/game-art-direction/agents/openai.yaml",
@@ -586,19 +586,24 @@
     },
     {
       "path": "skills/game-art-direction/references/art-direction-method.md",
-      "sha256": "8e9879bae6cedc01047792d305154e6bd9419ec9c455c541162b372be7395bdd",
-      "bytes": 2811
+      "sha256": "367b8d421b99c64151582f3803b9b07689ded244b32125cb50c86cf49cb19f41",
+      "bytes": 3162
     },
     {
       "path": "skills/game-build/SKILL.md",
-      "sha256": "a3b01dfc8dff5cb06e57699df23a7db96b38eec709aae11ef3d90f78acd5915d",
-      "bytes": 4678
+      "sha256": "4a1b79ed036d2f149fb0a19f68609bf7fd0a5335a171321b7bd67175cd6671d2",
+      "bytes": 5563
     },
     {
       "path": "skills/game-build/agents/openai.yaml",
       "sha256": "4fea64c3910933415a5e601bc3f423d22f0e5cfed8d7ab5c2aa38f60a3105cf0",
       "bytes": 272
     },
+    {
+      "path": "skills/game-build/references/baked-sprite-workflow.md",
+      "sha256": "aa7cbf2c255bae747a65b4d8eaef060ec77a4e3da137636d090e1b4583e390a5",
+      "bytes": 1671
+    },
     {
       "path": "skills/game-build/references/blender-asset-workflow.md",
       "sha256": "6fbf2e11f74ca2be1d2e3d3644dfc9ed82f04a6f0d92b67b2267894a710a4a42",
@@ -626,7 +631,7 @@
     },
     {
       "path": "skills/game-concept/SKILL.md",
-      "sha256": "fd07b5f61f4408887e0fb982c74edb42735bf25c456d6ca2dad60dff3f19b0ad",
+      "sha256": "bce6fadcf21c57ce279dcb90af6f3d15ea9328410f291977a1c9cae4d3d6de20",
       "bytes": 3436
     },
     {
@@ -636,8 +641,8 @@
     },
     {
       "path": "skills/game-concept/references/concept-method.md",
-      "sha256": "ae45e4ffafe6c253e66ea06636d836f81485bf0b705fc3e6fbb4be8f94a0c859",
-      "bytes": 2939
+      "sha256": "c34c47b3920740957dbf4b8e64c7258e1b584604eff671ca2718baf340bc9538",
+      "bytes": 3284
     },
     {
       "path": "skills/game-qa/SKILL.md",
@@ -656,18 +661,18 @@
     },
     {
       "path": "skills/game-qa/references/test-design-method.md",
-      "sha256": "f0f52f7b791d59fb55942257a53b69367c01eadf57dd3f918f4f4482cec478ca",
-      "bytes": 3770
+      "sha256": "fa6597233fed41464bec16b9253c54357d2391a1ca5791b223049a7e2a3dad9c",
+      "bytes": 3771
     },
     {
       "path": "skills/game-world-design/SKILL.md",
-      "sha256": "3d83f142305906afa9174adba07b7bde2c163489cf4daa7118bc2a113e818c6f",
-      "bytes": 3883
+      "sha256": "c72856e53a0ec22e0c33c2f9fb593e426cdf014d319aecd949c0be25518c693c",
+      "bytes": 3913
     },
     {
       "path": "skills/game-world-design/agents/openai.yaml",
-      "sha256": "4447bb4f6905cc367d6bfc45827877e16c5b3bf4b3ddc00fd3547fe31397bebd",
-      "bytes": 280
+      "sha256": "7dd8700b278aa96b4cf68e5c394c958d788a45a8041c8fe4fbd3b5ac817965bf",
+      "bytes": 336
     },
     {
       "path": "skills/game-world-design/references/narrative-design-method.md",
@@ -701,8 +706,8 @@
     },
     {
       "path": "skills/novel-to-game/SKILL.md",
-      "sha256": "12c82429d8bfeff5fa12bc839c618ef03897c97cb1126e0002236c59b3837957",
-      "bytes": 4223
+      "sha256": "b8ed42dd6d6e2d0a38f8f4014948b3d5a35f95199e364054878a256413105258",
+      "bytes": 4485
     },
     {
       "path": "skills/novel-to-game/agents/openai.yaml",
@@ -711,13 +716,13 @@
     },
     {
       "path": "skills/novel-to-game/references/intake-method.md",
-      "sha256": "a704a8d9132e94bd425c778d62a2db3d74cc48a1e16de981a9b18d3bf66c3f17",
-      "bytes": 3004
+      "sha256": "6ded64de535ed03f996e336fcc413adf358b454cc2b3e04331d83400b7daffae",
+      "bytes": 3112
     },
     {
       "path": "skills/novel-to-game/references/pipeline-contract.md",
-      "sha256": "96375b18d8babc7f73ca429c64ef18bf1607d7058f92109fdc4002531b22cfa7",
-      "bytes": 2942
+      "sha256": "ce943fec9b9de37914b4ae949b76ce82e774f154ec0ac5a63e6fae6b2044962d",
+      "bytes": 3998
     }
   ]
 }

package/lib/novel-to-game/package.json

@@ -0,0 +1,4 @@
+{
+  "private": true,
+  "description": "Bundled upstream Skills; no module type so scripts resolve as in an upstream install."
+}

package/lib/novel-to-game/skills/game-art-direction/SKILL.md

@@ -17,9 +17,9 @@ description: "Direct game art and creative vision. Turn GAME_DESIGN into ART_DIR
 
 ## 设计
 
-1. 定义能区分正确与错误方案的视觉原则；身份与审美校准按方法同名一节；
+1. 定义能区分正确与错误方案的视觉原则；身份与审美校准见参考方法同名一节；
 2. 为玩家、目标、威胁、奖励和地标建立镜头、焦点、轮廓、尺度与密度语法；
-3. 让颜色、光、材质和运动编码真实状态，并提供非颜色冗余、低动效表达与明度可读边界；
+3. 让颜色、光、材质和运动编码真实状态，并提供不依赖颜色的提示、低动效表达与清晰的明度对比；
 4. 定义首屏焦点、HUD 信息层级、输入反馈、结果分级、失败与结果状态；
 5. 为环境、角色、交互物和关键道具定义共享世界语法与各自识别锚点；
 6. 定义目标语言的字体覆盖、阅读顺序、文本密度与文化符号边界；

package/lib/novel-to-game/skills/game-art-direction/references/art-direction-method.md

@@ -31,6 +31,10 @@
 仅在高风险歧义或目标成色需要时制作视觉目标包：原创风格帧、构图草图或白盒改绘，
 附能裁决方向的可观察条件和失败例。应包含实际游戏界面，不用海报证明可操作性。
 
+每个身份锚点、关键时刻和主要界面先出约三张方向可区分的原创候选。流程处于 `director` 模式时停靠，
+由用户为每一项选定一张；`quick` 模式按这些可观察条件自选并写明理由。选中图进入视觉目标包，成为批量
+生产唯一的参考；落选图与被否决样张一样隔离。
+
 批准目标图后批量资产保持身份、光色、构图与文本空间；方向被否决时停止同方向扩批，
 隔离被否决样张，不让旧图继续作为生产参考。目标图只锁方向，不证明最终资产或发布质量。
 工具不可用时保留文字方向并记录未解决风险，不伪称图像已完成。

package/lib/novel-to-game/skills/game-build/SKILL.md

@@ -37,6 +37,8 @@ BUILD_BRIEF 只压缩产品边界、必须保真的体验事实、运行方式
   [playable-model-contract.md](references/playable-model-contract.md) 的相关部分。
 - 语音策略不是 `none` 时读取 [tts-production-contract.md](references/tts-production-contract.md)。
 - 实际采用动态媒体时读取 [generative-media-pipeline.md](references/generative-media-pipeline.md)。
+- 2.5D 角色或物件需要多方向、多动作且身份一致的序列帧时，读取
+  [baked-sprite-workflow.md](references/baked-sprite-workflow.md)。
 
 ## 共同构建循环
 
@@ -45,7 +47,11 @@ BUILD_BRIEF 只压缩产品边界、必须保真的体验事实、运行方式
 2. 实现 GAME_DESIGN 的状态、动作与结果；使用生成模型时，模型只解释输入和结果，规则器拥有提交权。
    存在 `signature_command` 时再执行可选合同的专属边界。
 3. 保留能复现问题的初态与输入路径，涉及随机性时固定 seed；修订后重放失败路径与相邻反例。
+   同一手感或表现问题连续两次修复无效时停止换方案，先查成熟作品处理同类问题的公开资料（开发者
+   演讲、技术复盘或开源实现），理解其原理后再实现。
 4. 回写实际工具链、install/build/start 命令和版本；未知值写 `NOT_AVAILABLE: 原因`，不猜。
+   实际候选与 GAME_DESIGN、ART_DIRECTION 的事实描述不符时，交回对应 owner 修订或在 BUILD_BRIEF 标明冲突，
+   不留下描述已不存在的玩法、表现或声音的文档。
 
 `whitebox` 到此按 build-brief-contract「完成证据」运行最窄检查并交回 design owner；设计修订后
 重放受影响路径，直到最大风险已被实际暴露或当前方向被否决。
@@ -57,7 +63,9 @@ BUILD_BRIEF 只压缩产品边界、必须保真的体验事实、运行方式
 6. 运行最窄的开发检查与启动 smoke，修复构建失败、阻断日志、资源失败和崩溃；替代运行时未覆盖的
    目标平台输入、性能、打包或设备项写入 limitation。
 7. 达到 brief 的 `targetFinish`；更高完成度只处理已批准的焦点资产和招牌时刻，不制造与可玩闭环
-   无关的发布审计。
+   无关的发布审计。有批准目标图的界面与焦点资产，在目标视口截取实际运行画面与目标图并排比对，
+   逐项列出布局、比例、字体、色光与层级差异后修正；连续两轮没有改善就停止，把剩余差异写进
+   BUILD_BRIEF 的限制，不修改目标图去迁就实现。
 8. 交给 `game-qa`；诊断、复跑与事实源规则见契约「完成证据」。
 
 ## 输出

package/lib/novel-to-game/skills/game-build/references/baked-sprite-workflow.md

@@ -0,0 +1,19 @@
+# 2.5D 序列帧烘焙工作流（按需）
+
+适用于 2.5D 游戏里需要多方向、多动作且身份一致的角色或物件。逐帧二维生成很难在转向和动作之间保持
+同一张脸、同一套服装和同一比例；先做出三维模型、再按游戏镜头逐帧渲染，可以把这些事实锁在模型里。
+已有可用序列帧或程序化表现已经足够时不迁移。它是资产制作路径，不替代目标运行时，也不改变
+GAME_DESIGN 与 ART_DIRECTION 的裁决。
+
+## 流程
+
+1. 以 ART_DIRECTION 批准的角色目标图为唯一输入生成三维模型，对照目标图核对脸、发型、服装、标志道具
+   与比例；识别锚点偏离时重新生成，漂移的模型不进入后续步骤。
+2. 自动绑定骨骼，按 GAME_DESIGN 实际用到的状态套用或制作动作，检查关节弯曲处是否穿模、拉伸。
+3. 在目标引擎的无头模式或三维软件中，按游戏实际的镜头俯角、投影方式、主光方向和描边规则，逐方向、
+   逐动作渲染；同一批资产统一方向数、帧率、画布尺寸和脚底锚点。
+4. 打包图集并配置资产键，放回实际游戏镜头检查：转向是否顿挫（方向数不足或缺转身过渡）、移动时脚步
+   是否打滑、锚点是否抖动、光色是否与场景一致。只重做暴露问题的方向或动作。
+
+工具、模型与服务按当前能力、许可与费用选择；上传内容只含该角色的目标图。源模型、绑定文件、渲染脚本
+和最终图集都保存在工作区，供应商任务页不算资产。必需资产生成失败时明确报错，不用占位图静默通过。

package/lib/novel-to-game/skills/game-concept/SKILL.md

@@ -44,7 +44,7 @@ description: "Design game concepts from a novel. From SOURCE_BIBLE and PRODUCT_B
 3. 紧凑先例说明；
 4. 候选概念卡：主类型、`experienceProfile`、借用/拒绝的玩法先例、核心动作与循环、压力、失败、
    世界响应；说明玩家身份如何形成反复行动、执行阻力与可观察变化，不要求独立的固定键名。
-   `signature_command` 按方法「身份专属命令」写值或 `N/A`；最后说明原作张力如何由玩家亲手 enact、
+   `signature_command` 按方法「身份专属命令」写值或 `N/A`；最后说明原作张力如何由玩家行动体现、
    招牌画面、最小验证切片、最大风险和否决观察；
 5. 硬否决结果、关键取舍、推荐与选择状态；
 6. 选定方向的玩家能动性合同、不可妥协体验承诺；

package/lib/novel-to-game/skills/game-concept/references/concept-method.md

@@ -45,5 +45,8 @@
 
 ## 选择依据
 
-先否决，再比较原作转化、玩家能动性、文化表达与最小验证成本。
+先否决，再比较原作转化、玩家能动性、文化表达、最小验证成本与内容成本。
+内容成本指每分钟游玩时长需要多少专门制作的场景、角色与文本。靠组合扩展内容的方向（装备词缀、可拼接的
+地图块、敌人变体、可复用的事件模板）比每章一套专属场景和演出更容易在预算内做完；小说改编天然偏向后者，
+所以要显式比较。
 推荐能用现有时间、工具和资产证明核心幻想、且否决观察最清楚的方向；相近方案不做伪精确排名。

package/lib/novel-to-game/skills/game-qa/references/test-design-method.md

@@ -12,7 +12,7 @@
 个正例和一个有区分力的边界或反例。断言盯住玩家可观察结果，不绑定内部函数名。
 
 叙事项目重点检查关键选择、持久状态读取、分支汇流、结局条件和人物知识边界；系统项目重点检查
-状态迁移、门槛、资源、失败和重开。修复缺陷时补能重现原故障的回归，不只验证 happy path。
+状态迁移、门槛、资源、失败和重开。修复缺陷时补能重现原故障的回归，不只验证成功路径。
 
 有事件日志时，保存规则版本、初态摘要、seed、输入序列和终态摘要，完整重放一次；随机探索只用于找
 死路和失败 seed，不能代替真实路径。叙事生成存在时增加四类反例：事实冲突、承诺冲突、改写玩家已

package/lib/novel-to-game/skills/game-world-design/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: game-world-design
-description: "Design game experience, systems, and levels. Converge the chosen concept into one GAME_DESIGN defining the player promise, core loop, how the world responds, the systems actually needed, level pacing, feedback, failure, and a fully playable prototype. Use for design the game world, deepen the gameplay and levels, write the game design document, design the branching story structure, scenes, dialogue and consequences. 游戏体验、系统与关卡策划。把选定概念收束为一份 GAME_DESIGN，定义玩家承诺、核心循环、世界响应、必要系统、关卡节奏、反馈、失败和可完整游玩的原型。用于设计游戏世界、深化玩法和关卡、设计互动叙事的场景对白与分支因果等需求。"
+description: "Design game experience, systems, and levels. Converge the chosen concept into one GAME_DESIGN defining the player promise, core loop, how the world responds, the systems actually needed, level pacing, feedback, failure, and the specification for a complete playable slice. Use for design the game world, deepen the gameplay and levels, write the game design document, design the branching story structure, scenes, dialogue and consequences. 游戏体验、系统与关卡策划。把选定概念收束为一份 GAME_DESIGN，定义玩家承诺、核心循环、世界响应、必要系统、关卡节奏、反馈、失败及可供构建的完整切片设计。用于设计游戏世界、深化玩法和关卡、设计互动叙事的场景对白与分支因果等需求。"
 ---
 # 游戏体验与世界设计
 

package/lib/novel-to-game/skills/game-world-design/agents/openai.yaml

@@ -1,4 +1,4 @@
 interface:
   display_name: "Game World Design"
   short_description: "Design the playable world, systems, and levels · 设计可玩的世界、系统与关卡体验"
-  default_prompt: "Use $game-world-design to deepen this concept into a playable world and a complete prototype."
+  default_prompt: "Use $game-world-design to write GAME_DESIGN for this concept, defining the world, systems, levels, and a complete playable slice for the build stage."

package/lib/novel-to-game/skills/novel-to-game/SKILL.md

@@ -21,7 +21,8 @@ description: "Turn a novel into a fully playable game on the selected target pla
 ## 模式
 
 - `quick`：默认；用推荐草案推进，比较有效替代后自动选择并完成全流程。
-- `director`：给出候选和推荐后停靠，等待用户选方向；已有明确选择时不重复停靠。
+- `director`：在概念方向、白盒试玩后的内容取舍和视觉目标候选三处给出候选与推荐后停靠，等待用户选择；
+  已有明确选择时不重复停靠。
 - `resume`：读取 `_progress.md` 和实际产物，从最早未完成的交接继续。
 
 ## 流程
@@ -33,8 +34,10 @@ description: "Turn a novel into a fully playable game on the selected target pla
    不适用时明确 `N/A`；`director` 的待决方向在此停靠。
 5. 调用 `game-world-design` 生成 `GAME_DESIGN.md`，其中锁定最大设计风险与最小可执行切片。
 6. 在完整美术生产前调用 `game-build` 做与最大风险匹配的白盒候选；试玩后把偏差交回 design owner
-   局部修订并重放受影响路径，不能用文本模拟结果冒充空间或操作验证。
-7. 调用 `game-art-direction` 生成 `ART_DIRECTION.md`；只有目标成色需要时再制作视觉目标包。
+   局部修订并重放受影响路径，不能用文本模拟结果冒充空间或操作验证；随后按流程契约做第一次预算与
+   范围复盘，据此请 design owner 调整剧情与关卡内容量。
+7. 调用 `game-art-direction` 生成 `ART_DIRECTION.md`；只有高风险歧义或目标成色需要时再制作视觉目标包，目标图选定后
+   才进入批量资产生产。
 8. 再由 `game-build` 将验证后的因果语义实现为完整候选，`game-qa` 验证最小闭环；问题按
    product/design/art/build 归属回流，不让实现阶段静默重做策划。
 

package/lib/novel-to-game/skills/novel-to-game/references/intake-method.md

@@ -1,7 +1,7 @@
-# 需求 intake：先给推荐，再处理高风险歧义
+# 需求确认：先给推荐，再处理高风险歧义
 
-intake 的目的不是举行逐项审批会，而是尽早锁住会让下游返工的产品边界。先读来源，替用户写一页
-推荐草案；低风险空白列为未确认假设并继续，只有 materially branching 或安全相关决定才停靠。
+尽早锁定会影响后续制作的产品边界。先读来源，替用户写一页推荐草案；低风险空白列为未确认假设并继续，
+只有会实质改变方向或涉及安全的决定才停下来确认。
 
 ## 默认动作
 
@@ -13,7 +13,7 @@ intake 的目的不是举行逐项审批会，而是尽早锁住会让下游返
    brief 显式修订。
 
 联网只用于会影响真实决定且可能变化的事实，例如当前引擎导出能力、许可证、渠道限制和新服务规格。
-不要为了给每个常识默认值找出处而把 intake 变成研究项目。
+只核实会影响决策的信息，无需为常识默认值逐项查找出处。
 
 ## 产品框架
 
@@ -31,6 +31,7 @@ intake 的目的不是举行逐项审批会，而是尽早锁住会让下游返
 | 内容尺度 | 分级、成人内容范围与平台约束 |
 | 核心幻想 | 玩家扮谁、追求什么、何时产生什么情绪 |
 | 时长与结构 | 验证切片时长、单局/长线结构 |
+| 制作预算 | 本次可用的模型额度、工时或费用上限；用户未给时按对话中的可用资源估计并列入未确认假设 |
 | 发行定位 | 买断、免费、N/A；只记录会影响玩法的部分 |
 | 引擎与运行时 | 生产引擎、目标运行时、可用的实际测试运行时 |
 | 玩家结构 | 单人/多人/异步/实时，以及切片内真实范围 |

package/lib/novel-to-game/skills/novel-to-game/references/pipeline-contract.md

@@ -32,7 +32,7 @@ JSON/YAML。设计理由仍由 `GAME_DESIGN.md` 拥有，运行状态与动作
 | `scope` | brief、source bible 和三份设计交接存在；范围、原作事实、目标运行形态、`targetFinish` 与 `experienceProfile` 不冲突 |
 | `playable` | `qa/verification.json` 的 `launch`、`render`、`input`、`coreLoop`、`outcome`、`restart` 均有真实运行证据 |
 
-`_progress.md` 只记录来源、模式、当前阶段、未确认假设、回流和这两项结果。详细测试状态留在
+`_progress.md` 只记录来源、模式、当前阶段、未确认假设、回流、预算复盘结论和这两项结果。详细测试状态留在
 `qa/verification.json`，不要复制到多份状态表。
 
 早期工作区缺少 `experienceProfile` 时，`resume` 按现有 `CONCEPT.md` 补记一次并继续；不得据此重做
@@ -41,12 +41,23 @@ JSON/YAML。设计理由仍由 `GAME_DESIGN.md` 拥有，运行状态与动作
 ## 证据角色
 
 构建准备候选与 verify 入口。QA 可诊断、修复后复跑；最终事实源原子记录当前同一次完整运行，
-不拼接旧 PASS。不要求真人试玩、逐项人工批准、重复证据对象或第二份 QA 报告。
+不拼接旧 PASS。QA 结论不要求真人试玩、逐项人工批准、重复证据对象或第二份 QA 报告。
 
 白盒试玩是设计回流，不是第七道顶层 QA 门。反馈至少记录观察位置、玩家输入、预期与实际差异、受影响
 的状态/动作/知识/回响节点和 owner；修订后用可复现的初态与输入路径复查，随机性相关时固定 seed。
 已有存档/日志因节点标识、变量语义、动作效果或知识权限变化而失效时，迁移或明确拒绝兼容，不能静默解释成新历史。
 
+白盒试玩还要回答玩家实际追着玩、反复回来做的是哪一部分。`director` 可在此停靠，请用户找一位目标
+玩家试玩；无人可找时，另起一个未读过设计文档与源码的代理会话，只凭画面和正常输入试玩。概念与设计阶段
+写下的剧情、对白和关卡内容预算按试玩结论修订，与核心乐趣无关的原作情节压成过场或删去。
+
+## 预算与范围复盘
+
+`PRODUCT_BRIEF.md` 记录本次制作可用的模型额度、工时或费用上限。白盒试玩结束、生产阶段完成约一半必做
+项时各复盘一次，把剩余预算与剩余范围对照；单次代理任务的耗时或消耗持续上升、并行代理反复改动同一批
+文件，都说明工程复杂度已逼近预算能支撑的上限。此时停止加内容，按玩家实际看到的收益删减或降级：做完
+少量内容胜过铺满占位，打磨先投给常驻屏幕的界面与主角，再到场景。每次删减写进 brief 的明确非目标。
+
 预算或工具耗尽只会留下 `NOT_RUN` / `FAIL`、缩小范围或延期，不会生成 PASS。
 
 ## resume 与回流