Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →使用 NovelWeave 进行小说创作的完整工作流程，包括命令使用、最佳实践和高效创作技巧。适用于规划小说项目、组织创作过程或学习 NovelWeave 功能。
.claude/skills/microck-novelweave-workflow/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 49% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -18% | 0% |
NovelWeave 是一个 AI 驱动的小说创作助手，设计用于支持从构思到完稿的整个创作过程。
使用 /constitution 命令建立小说的核心价值观和主题。
目的：
示例：
/constitution
主题：救赎与希望
价值观：即使在最黑暗的时刻，人性的善良仍会发光
避免：廉价的情感操纵、不必要的暴力使用 /plan 命令构建小说大纲和结构。
最佳实践：
使用 Agent Rules 和 Knowledge Base 创建详细的角色档案。
关键要素：
为奇幻、科幻或复杂背景建立世界观。
记录内容：
/write 命令这是核心创作命令，用于生成场景内容。
有效的 /write 请求：
/write
场景：艾米在废弃工厂与追踪者对峙
情感：紧张、恐惧但决心坚定
重点：展示艾米的机智和勇气
长度：800-1000 字避免：
使用 /track 系统维护情节、角色和时间线的一致性。
功能：
使用一致性检查工具审查稿件。
检查领域：
审查对话、节奏和描写质量。
常见改进点：
❌ 过度依赖 AI - 你是作者，AI 是助手 ❌ 跳过规划 - 一些结构能防止后期大返工 ❌ 忽视一致性 - 小错误会累积破坏可信度 ❌ 一次写太多 - 专注质量而非数量
/constitution - 定义核心价值观和主题/plan - 创建和管理小说大纲/write - 生成场景内容/track - 追踪情节和角色一致性/clarify - 解决情节问题和填补漏洞/analyze - 深入分析文本和结构记住：NovelWeave 旨在增强你的创造力，而非替代它。保持你的独特声音，让 AI 帮助你实现愿景。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 20,935 | 10,733 | -49% | 1 | 1 | 0% | 3,063 | 2,803 | -8% | 0 | 0 | — |
case-02 | fail→pass | 13,668 | 12,846 | -6% | 1 | 1 | 0% | 1,987 | 2,954 | +49% | 0 | 0 | — |
case-11 | pass→pass | 19,555 | 7,337 | -62% | 1 | 1 | 0% | 2,797 | 2,248 | -20% | 0 | 0 | — |
case-03 | fail→fail | 22,113 | 19,399 | -12% | 1 | 1 | 0% | 2,621 | 3,658 | +40% | 0 | 0 | — |
case-04 | fail→pass | 17,320 | 9,534 | -45% | 1 | 1 | 0% | 2,316 | 2,476 | +7% | 0 | 0 | — |
case-05 | fail→pass | 14,847 | 8,395 | -43% | 1 | 1 | 0% | 2,132 | 2,344 | +10% | 0 | 0 | — |
case-06 | fail→pass | 18,313 | 7,400 | -60% | 1 | 1 | 0% | 2,620 | 2,140 | -18% | 0 | 0 | — |
case-07 | pass→pass | 15,955 | 11,638 | -27% | 1 | 1 | 0% | 2,338 | 2,994 | +28% | 0 | 0 | — |
case-08 | pass→pass | 17,012 | 10,360 | -39% | 1 | 1 | 0% | 2,478 | 2,746 | +11% | 0 | 0 | — |
case-09 | pass→pass | 20,771 | 18,326 | -12% | 1 | 1 | 0% | 2,793 | 3,696 | +32% | 0 | 0 | — |
case-10 | fail→pass | 18,796 | 16,744 | -11% | 1 | 1 | 0% | 2,522 | 3,259 | +29% | 0 | 0 | — |
case-12 | pass→pass | 21,765 | 20,808 | -4% | 1 | 1 | 0% | 3,045 | 4,347 | +43% | 0 | 0 | — |
case-13 | fail→pass | 9,251 | 3,749 | -59% | 1 | 1 | 0% | 1,424 | 1,737 | +22% | 0 | 0 | — |
case-14 | pass→pass | 16,673 | 25,348 | +52% | 1 | 1 | 0% | 2,324 | 3,548 | +53% | 0 | 0 | — |
case-15 | fail→pass | 15,648 | 7,598 | -51% | 1 | 1 | 0% | 2,229 | 2,277 | +2% | 0 | 0 | — |
case-16 | pass→pass | 18,621 | 17,528 | -6% | 1 | 1 | 0% | 2,688 | 3,550 | +32% | 0 | 0 | — |
case-17 | pass→pass | 17,113 | 10,880 | -36% | 1 | 1 | 0% | 2,406 | 2,742 | +14% | 0 | 0 | — |
case-18 | pass→pass | 16,964 | 12,636 | -26% | 1 | 1 | 0% | 2,376 | 2,848 | +20% | 0 | 0 | — |
case-19 | pass→pass | 16,167 | 7,571 | -53% | 1 | 1 | 0% | 2,169 | 2,256 | +4% | 0 | 0 | — |
case-20 | pass→pass | 20,626 | 19,724 | -4% | 1 | 1 | 0% | 3,035 | 3,964 | +31% | 0 | 0 | — |
case-21 | pass→pass | 19,535 | 18,777 | -4% | 1 | 1 | 0% | 2,805 | 3,932 | +40% | 0 | 0 | — |
case-22 | pass→pass | 19,573 | 19,425 | -1% | 1 | 1 | 0% | 3,195 | 3,654 | +14% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.