inyhow-ai-novel-generator-web-novel-workflow.ocm.json json {
"ocm": "1",
"id": "inyhow-ai-novel-generator-web-novel-workflow",
"kind": "skill",
"name": "web-novel-workflow",
"description": "“Structured workflow for Chinese long-form web novel creation (中文长篇网文). Guides 5-question confirmation, chapter planning with 3000-5000 word targets, suspense hook design, and de-AI polishing. Use when a user asks to write a novel (写小说), create web fiction (网络小说/连载小说), plan chapter outlines (章节规划), expand drafts (扩写), or reduce AI-sounding prose (去AI化).”",
"publisher": "inyhow",
"version": "1.0.0",
"capabilities": {
"domains": [
"creative",
"math"
],
"tags": [
"skill-md",
"github"
],
"languages": [
"en"
]
},
"quality_prior": 0.6,
"examples": [
"“Structured workflow for Chinese long-form web novel creation (中文长篇网文). Guides 5-question confirmation, chapter planning with 3000-5000 word targets, suspense hook design, and de-AI polishing. Use when a user asks to write a novel (写小说), create web fiction (网络小说/连载小说), plan chapter outlines (章节规划), expand drafts (扩写), or reduce AI-sounding prose (去AI化).”"
],
"primary": false,
"metadata": {
"source": {
"provider": "github",
"repository": "https://github.com/inyhow/ai-novel-generator",
"path": "SKILL.md",
"ref": "7e5842c4df6661fb7cfaecf472c7215d2a955be4",
"url": "https://github.com/inyhow/ai-novel-generator/blob/7e5842c4df6661fb7cfaecf472c7215d2a955be4/SKILL.md",
"key": "inyhow/ai-novel-generator/SKILL.md"
}
},
"instructions": "# 中文网文创作工作流\n\n## 触发场景\n\n- 用户要求写小说、创作长篇、生成网络小说或连载小说。\n- 用户要求按固定步骤做创作拆解和章节推进。\n- 用户要求扩写现有章节或草稿。\n- 用户要求提升”人味”、降低AI痕迹（去AI化润色）。\n\n## 快速流程\n\n1. 执行5问确认 → 用户确认后才继续。\n2. 生成总纲与角色弧线 → 用户审阅并确认。\n3. 按10-50章进行章节创作（单章目标3000-5000字）。\n4. 每章结尾补悬念钩子。\n5. 每5章检查点：验证角色动机连续性、主线冲突推进、前后因果连贯。\n6. 做去AI痕迹润色。\n\n## 5问确认\n\n在创作前必须收集以下信息，缺一不可：\n\n1. 题材与风格是什么？（如：东方玄幻、都市言情、悬疑推理）\n2. 主角身份、欲望和核心缺陷是什么？\n3. 世界设定与主冲突是什么？\n4. 节奏与叙事语气偏好是什么？（如：紧凑热血、慢热细腻）\n5. 篇幅目标（章数、单章字数）是什么？\n\n## 输出规范\n\n- 标题使用《》\n- 先给150-300字简介\n- 正文按”第X章 章节名”输出\n- 保证角色动机一致、前后因果连贯\n- 避免模板化总结句和机械复读\n- 参考 `references/writing-standards.md` 了解质量标准\n- 参考 `references/sensitive_words.txt` 避免敏感词\n\n## 悬念",
"cost": {
"context_tokens": 260
}
}
