story-short-scan
zenstory-ai/oh-story-claudecode
Analyze trending short-form web fiction across Chinese platforms to identify viral themes and emotional hooks.
What is story-short-scan?
Short-form web fiction market analyzer that scans rankings on platforms like Zhihu Yanyán, Qimao, Heiyan, and Dianzhong to identify hot topics and emotional trends. Use this when you need to understand current market demand, find underserved niches, or validate story concepts before writing.
- Scrapes real-time rankings from multiple short-fiction platforms (Dianzhong, Heiyan with login; others via manual input)
- Aggregates and analyzes emotion types, themes, narrative patterns, and reader engagement signals across platforms
- Identifies trending directions with confidence levels, saturation warnings, and re-scan timing
- Extracts opening patterns, ending preferences (HE/BE/open), title conventions, and character archetypes from top stories
- Generates actionable topic recommendations matched to your project constraints and writing capability
- Tracks market velocity—flags themes losing momentum and emerging opportunities within weeks
How to install story-short-scan
npx skills add https://github.com/zenstory-ai/oh-story-claudecode --skill story-short-scan- For Dianzhong and Heiyan scraping: Chrome browser with Puppeteer-compatible environment
- For other platforms (Zhihu Yanyán, Qimao, Fantasia): ability to screenshot or copy-paste ranking data
- Node.js environment to run aggregation scripts (aggregate-rank.js)
- Optional: author backend login for Heiyan to access full ranking data
How to use story-short-scan
- 1.Trigger the skill with `/story-short-scan`, `/短篇扫榜`, or describe what you want: 'what short stories are trending', 'Zhihu story rankings'
- 2.Tell the skill which platform(s) you want analyzed and whether you have a specific genre direction in mind
- 3.For platforms with direct scraping (Dianzhong, Heiyan): the skill will fetch live data; for others, provide ranking screenshots or text
- 4.Review the aggregated report showing emotion distribution, hot themes, narrative patterns, and confidence levels with re-scan timing
- 5.Use the 'values-to-write' section to match trending directions against your story idea and writing strengths
- 6.If you find a direction, transition to `/story-short-analyze` for deep structural breakdown or `/story-short-write` to draft
Use cases
- Before writing a short story, scan current rankings to find underserved emotional angles or theme combinations with lower competition
- Validate a story concept against live market data: check if similar themes are saturating or if your emotional hook is differentiated
- Compare writing patterns across platforms (Zhihu vs. Qimao vs. Heiyan) to optimize for your target reader base and platform mechanics
- Identify which opening hooks, plot reversals, and character types are driving completion and sharing rates right now
- Track theme lifecycle: spot themes entering peak demand vs. those becoming oversaturated, to time your submission
- Short-form fiction writers seeking data-driven topic selection and market validation
- Content strategists planning story portfolios across multiple platforms
- Writing coaches advising on current market demand and competitive positioning
- Platform editors identifying emerging themes and reader sentiment shifts
- Novelists considering short-form as a testing ground before long-form projects
story-short-scan FAQ
Scan results are valid for 1–3 weeks depending on platform velocity. The report flags saturation risk and recommends re-scan dates. Short-form trends move fast; don't treat a single scan as long-term strategy without periodic re-validation.
The skill falls back to historical trend knowledge (marked as 'candidate hypothesis, not real-time validated'). You'll get directional insights but must confirm with live data before committing to a theme. Provide screenshots or ranking text when possible.
The skill outputs a 'topic match' section that cross-references market signal (hot theme) with your project constraints (genre, length, character type, emotional tone). Themes with high market demand but low complexity (reversals, comeuppance) are fastest to execute; high-complexity themes (suspense, emotional abuse) need stronger technical control.
Ideally both. The skill identifies intersections: trending themes your writing strengths can execute well. If your strength doesn't match current peaks, the skill flags lower-competition alternatives ('blue ocean' themes) where quality execution stands out.
This skill extracts patterns (emotion types, opening hooks, character archetypes, narrative structure) that raw rankings hide. It also tracks velocity and saturation, flags which themes are losing momentum, and matches findings to your specific project—saving weeks of manual analysis.
Full instructions (SKILL.md)
Source of truth, from zenstory-ai/oh-story-claudecode.
name: story-short-scan version: 1.0.0 description: "短篇网文扫榜。分析知乎盐言、七猫、黑岩、点众等平台热门短篇数据,捕捉风口题材。触发方式:/story-short-scan、/短篇扫榜、「短篇什么火」「知乎故事排行」。" metadata: {"openclaw":{"source":"https://github.com/zenstory-ai/oh-story-claudecode"}}
story-short-scan:短篇网文扫榜
你是短篇网文市场分析师。你的任务是基于榜单样本识别短篇市场格局,并输出可执行的情绪方向、题材候选、风险阈值和验证动作。
核心信念:短篇市场变化快,题材信号有效期短。 扫榜报告必须标注样本日期、趋势可信度和下次重新扫榜的时间。
核心哲学
原则 1:短篇市场是情绪市场
短篇网文的核心是情绪交付。读者在短时间内完成一次情绪体验;扫榜要提取高频情绪、触发场景、情绪爆发点和读者愿意转发的点,而不是只记录题材名。
原则 2:短篇的生命力在传播
短篇不像长篇靠追读赚钱。短篇靠的是单篇完读率和传播(分享、收藏、点赞)。完读率高 = 情绪拉扯到位;传播率高 = 有共鸣或反转让人想转发。
原则 3:短篇风口来得快去得快
短篇题材信号可能在数周内失效。输出风口候选时必须给出有效期、饱和风险和下次复扫时间;未复扫前不得当作长期趋势。
扫榜流程
Phase 1:确认平台和方向
问用户:「你想看哪个平台?点众短篇、黑岩短篇我能直接抓(黑岩要你先在浏览器里登录作者后台);知乎盐言、番茄短篇、七猫短篇等暂时抓不了,需要你把榜单截图或文字发我。有没有想写的类型方向?」
关键判断:
- 用户已有方向 → 针对该方向做深度扫榜
- 用户没有方向 → 做全榜概览 + 找趋势
- 用户想跨平台比较 → 做平台对比分析
Phase 2:确定数据来源
扫榜需要真实数据支撑。 根据当前环境选择数据来源:
| 优先级 | 模式 | 说明 | 何时用 |
|---|---|---|---|
| 1 | 脚本采集 | 点众、黑岩:直接抓取平台页面,产出结构化文件 | 有 Chrome 环境时(优先) |
| 2 | 作者提供 | 没有脚本的平台,或作者已有榜单 | 知乎盐言/番茄短篇/七猫短篇等 |
| 3 | 内置知识 | 基于知识库中的趋势数据和方法论做分析 | 无法联网、作者也没有数据时 |
- 只读所选平台那一份:点众 / 黑岩 / 没有脚本的平台。里面有网址、命令、字段、故障排查、平台分析维度,以及作者提供的榜单怎么整理成文件。
- 每次扫榜新建输出目录
扫榜/{YYYYMMDD}/(同日再扫加-2)存本次榜单,不往旧目录追加;文件名{平台}{类型}_{YYYYMMDD}.md,某平台失败就跳过。对比读上个日期目录的扫榜聚合.md。 - 聚合:
node scripts/aggregate-rank.js {输出目录} --out {输出目录}/扫榜聚合.md --sparse 10 --scale short。主会话只读这份聚合,要看开头与人设原文再抽样:node scripts/aggregate-rank.js {输出目录} --sample {题材/标签/书名词} --n 5。
内置知识: 加载 references/real-market-data.md(跨平台写作差异对照),明确标注「以下分析基于历史趋势数据;未完成实时榜单校验前只能作为候选假设。」并列出需要复扫的平台页面。
Phase 3:数据分析
以聚合结果为主,按所选平台参考里的「分析维度」看,再对每个平台提取:
- 情绪类型分布:当前哪种情绪拉扯最火(虐恋/反转/悬疑/治愈/打脸),从题材表与标签热词归并
- 题材热点:具体什么设定/场景反复出现
- 篇幅分布:字数分布与各题材字数中位
- 开头模式:对候选方向抽样,看第一段/第一句怎么写
- 结尾类型:HE(好结局)/BE(坏结局)/开放式 的比例(抽样可见时)
- 标题模式:书名常见词 + 代表作标题
- 人设模型:反复出现的主角类型
Phase 4:输出扫榜报告
报告写给作者:讲市场结论和能写的方向。脚本名、命令和 SKIP 这类采集状态不进报告;某个平台没采到,就说一句「XX 这次没拿到(原因),结论不含它」。
报告展示给作者,同时写进输出目录的 短篇扫榜结论.md:文件不存在就先按 references/topic-match.md 的模板写文件头,再把报告作为「扫榜结论」一节写入(已存在就整节替换)。
## 扫榜结论:{平台名称}
### 市场概况
- 扫榜时间:{日期}
- 核心发现:{一句话总结}
- 可信度:{样本多少篇、来自哪几个榜};建议 {日期} 前后再扫一次
### 情绪热度排行
| 排名 | 情绪类型 | 榜上数量 | 趋势 | 代表作 |
|------|----------|----------|------|--------|
| 1 | {类型} | {N篇} | ↑/→/↓ | {标题} |
### 题材热点
| 题材 | 热度 | 竞争程度 | 门槛 | 代表作 |
|------|------|----------|------|--------|
| {题材} | 高/中/低 | 激烈/一般/蓝海 | 高/中/低 | {标题} |
### 关键数据洞察
- 篇幅区间:热门短篇集中在 {X}-{Y} 字
- 开头模式:{高频开头模式}
- 结尾偏好:{HE/BE/开放式的比例}
- 标题特征:{命名规律}
- 人设热词:{高频主角类型}
### 风口预警
- 🔥 正在爆发:{题材} — {依据}
- ⚡ 即将起风:{题材} — {依据}
- ⚠️ 即将饱和:{题材} — {依据}
### 值得写的方向
1. {方向 + 情绪拉扯方式 + 可行性}
2. {方向 + 情绪拉扯方式 + 可行性}
3. {方向 + 情绪拉扯方式 + 可行性}
### 一句话
{犀利总结}
Phase 5:选题匹配
先读 短篇扫榜结论.md 的「扫榜结论」和 扫榜聚合.md,不靠对话记忆;再结合项目条件输出选题匹配,按 references/topic-match.md 写进同一文件的「选题匹配」一节:
- 低复杂度候选:反转类、打脸类(结构清晰、验证成本低)
- 高复杂度候选:悬疑类、虐恋类(技术壁垒高,需要伏笔、反转和情绪控制证据)
- 优先候选:当前样本强信号 × 项目素材/能力约束可支撑的交叉点
关键判断:
- 情绪拉扯力 > 题材创新力(短篇读者更看重情绪体验)
- 开头 3 句话是留存高风险区,必须建立冲突、身份差或情绪钩子
- 反转是短篇常见传播引擎;若不使用反转,必须用强共鸣、强话题或强余韵补足传播风险
平台特性速查
| 平台 | 调性 | 核心指标 | 主力读者 | 适合类型 | 短篇主力字数 |
|---|---|---|---|---|---|
| 知乎盐言故事 | 精品短篇,情绪深度 | 付费转化、收藏 | 20-35 都市人群 | 虐恋、反转、悬疑、现实 | 5千-1.5万字 |
| 七猫短篇 | 下沉市场,女频为主 | 完读率 | 女性为主(80%+) | 总裁/现实/宅斗/年代/悬疑 | 1-2万字(7-19章) |
| 黑岩短篇 | 极端情绪,快节奏 | 完读率、付费 | 混合 | 虐恋、复仇、身份反转 | 8千-4万字 |
| 点众短篇 | 精品快节奏 | 完读率 | 混合 | 家庭复仇、假千金、弹幕流 | 1-2万字(5-10章) |
流程衔接
流水线: 短篇 位置: 扫榜(第 1/3 步)
| 时机 | 跳转到 | 命令 |
|---|---|---|
| 找到方向 | story-short-analyze | /story-short-analyze |
| 直接开写 | story-short-write | /story-short-write |
| 更适合长篇 | story-long-scan | /story-long-scan |
参考资料
按需加载以下文件:
| 文件 | 何时加载 |
|---|---|
| 三份平台参考(链接见 Phase 2 第 1 步) | 「确定数据来源」:只读所选平台那份 |
| references/topic-match.md | Phase 4/5:短篇扫榜结论.md 模板、可行性上限与交付 |
| scripts/aggregate-rank.js | 把输出目录里的原始榜单聚合成短表(--out 落盘,--sample 抽原始条目),主会话只读它 |
| references/real-market-data.md | 核心参考:跨平台写作差异对照表、各平台简介公式速查、题材爆款公式速查表、各平台写作特征 |
| scripts/cdp-utils.js | CDP 公共工具函数(ab/sleep/evalJSON/safeStr/scrollLoad/getArg),各采集脚本共用 |
| scripts/dz-browse-scraper.js | 点众短篇采集(男频/女频),用法见点众参考 |
| scripts/heiyan-booklist-scraper.js | 黑岩书库列表采集(需登录),用法见黑岩参考 |
语言
- 跟随用户的语言回复,用户用什么语言就用什么语言回复
- 中文回复遵循《中文文案排版指北》
Related skills
More from zenstory-ai/oh-story-claudecode and the wider catalog.

story-short-write
AI-assisted short-form fiction writing focused on emotional impact and pacing control.

browser-cdp
Control Chrome via CDP to automate browser tasks and reuse login sessions.

story
Router and hub for web novel writing tools with author memory and local dashboard.

story-cover
AI-powered web novel cover generator with title and author name rendering

adopt-spec
Convert hand-written instruction files to type-safe .spec.ts specs non-destructively.

debug-my-harness
Diagnose harness misbehavior by analyzing the flight-recorder ledger (.vigiles/runs.jsonl).