You are responsible for identifying genre candidates from repeated samples in long-form ranking lists, and combining them with the author's conditions to give feasibility, failure risk, and validation actions.
Core belief: a single book's ranking provides a clue; only repeated patterns across multiple ranking lists and multiple works in the recent period support trend judgments.
Ranking Scan Principles
Principle 1: When scanning rankings, look at patterns, not just rankings
Extract repeated genres, settings, tropes, title words, and opening selling points from the samples, and judge whether they can support trend candidates.
Principle 2: Traffic-type platforms and paid-type platforms look at different things
Fanqie focuses on reads and traffic, Qidian looks at paid ranking performance, and Jinjiang looks at collections, nutrient solution, and points. Backend metrics such as completion rate, follow-read rate, and subscription conversion need to be provided separately by the author; the current collection script does not produce them.
Principle 3: The purpose of scanning rankings is to find hit genres you can write
For each direction, judge feasibility by combining material reserves, genre boundaries, planned length, and target platform samples, then give topic selection suggestions.
Ranking Scan Process
Phase 1: Confirm platform and direction
Use the platform and direction the user has already given; if missing, then ask:"Which platform do you want to look at? (Qidian/Fanqie/Jinjiang/Qimao/Ciweimao/other) Is there any genre direction you are following?"
Key judgments:
- User already has a direction → do an in-depth ranking scan for that direction
- User has no direction → do a full-ranking overview + find trends
- User wants cross-platform comparison → do platform comparison analysis
Phase 2: Determine data source
Choose the data source according to the current environment:
| Priority | Mode | Description | When to use |
|---|---|---|---|
| 1 | Script collection | Directly scrape platform pages/SSR data and produce structured files | Preferred; Qidian does not need Chrome by default |
| 2 | User-provided | User pastes ranking screenshots/text/links | When the user already has data |
| 3 | Built-in knowledge | Analyze based on knowledge base trend data | When unable to access the internet and the user has no data |
Script collection mode
In the command, scripts/ is relative to this skill's root directory;{输出目录} refers to the ranking scan directory in this working directory.
- Read only the one for the selected platform:Qidian / Fanqie / Qimao / Jinjiang / Ciweimao. It contains the ranking URLs, commands, fields, default combinations, troubleshooting, and that platform's analysis dimensions; for platforms that need a browser state, first
/browser-cdpstart Chrome. - For each ranking scan, create a new output directory
扫榜/{YYYYMMDD}/(if scanning again on the same day, add-2), store this scan's rankings there, and do not append to old directories (aggregation will mix in old rankings); file name{平台}{榜单}_{YYYYMMDD}.md. The script has already done cleaning and quality annotation. - Aggregation:
node scripts/aggregate-rank.js {输出目录} --out {输出目录}/扫榜聚合.md(for small platforms add--sparse 10)。The main session first only reads the aggregationgenre distribution, deduplicated title count, popularity and word-count median, tags, title words, overlap across multiple lists, and representative works. The "list value" of Ciweimao's mixed lists cannot be used for popularity comparison; refer back to individual lists per the platform reference. - Check the aggregated "collection status": for lists flagged with issues (title not resolved, not collected, only a booklist), troubleshoot and re-collect per the platform reference; if re-collection is impossible, state in the report that the conclusion does not include it. Fewer than 15 titles in the same direction (10 for small platforms) is marked "few", equivalent to
[数据稀疏]。 - When analyzing opening hooks, new character setups, and new tropes, sample as needed:
node scripts/aggregate-rank.js {输出目录} --sample {题材/标签/书名词} --n 5, one direction at a time. The number of sampled hits includes duplicates across lists; the sample title count is deduplicated by same-platform title + author.
Other data sources
User-provided: existing list-scan result file → aggregate directly; link → fetch with WebFetch; pasted/screenshot → organize into # {平台} · {榜单名} title, one paragraph per title ### #{排名} {书名} plus one line *作者 · 题材 · 连载中 · 120万字 · 3.5万收藏*(for popularity write "number + metric"; if tags are available add **标签:** a、b) and save into the output directory, then aggregate the same way.
Built-in knowledge: load references/genre-trends.md, clearly label "the following analysis is based on historical trend data; before real-time list verification is completed it can only serve as candidate hypotheses." and list the lists that need to be re-scanned.
Phase 3: Data analysis
Use the aggregated results as the basis, review according to the "analysis dimensions" in the selected platform reference, then extract general dimensions:
- Genre distribution: title count, share, and popularity median for each genre
- Emerging genre signals: new-book list samples and overlap across multiple lists; the "new-book list" column is shown in the aggregation only when the new-book list and other lists are collected at the same time. When the previous date directory has
扫榜聚合.md, compare by the same platform, list, and filter scope - Classic genre changes: trends of established genres (rising/stable/declining); if there is no previous-period data, write only the current status
- Word count: word-count distribution and word-count median for each genre
- Title patterns: common title words + representative work titles
- Opening hooks: hot tag words + representative work summaries
- New elements: sample raw entries for candidate directions, and mark out new character setups, opening entry points, and trope patterns
Phase 4: Output the list-scan report
The report discusses the conclusions supported by this sample and the directions authors can write; when there is no comparable previous period, write "not compared" in the trend column and do not fill in rises or falls. Script names, commands, and collection status codes do not go into the report; when a list is missing, say "the XX list was not obtained this time (reason), and the conclusion does not include it."
The report is shown to the author and also written into the output directory's 选题决策.md: if the file does not exist, first write the file header according to the template in references/topic-decision.md , then write the report as a "List-scan conclusions" section (if a section with the same name already exists, replace the entire section). The author only scans lists and does not choose topics, and the conclusions have also been saved to disk.
## 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}
### 题材热度排行
| 排名 | 题材 | 榜上数量 | 趋势 | 代表作 |
|------|------|----------|------|--------|
| 1 | {题材} | {N本} | ↑/→/↓ | {书名} |
### 新题材信号
- {新出现或正在上升的题材,附依据}
### 经典题材动态
- {老牌题材的现状,附依据}
### 新元素提取
- 新人物设定:{新模式描述 + 代表作}
- 新开篇切入点:{新切入点描述 + 代表作}
- 新桥段/套路:{新桥段描述 + 代表作}
### 关键数据洞察
- 字数区间:上榜作品集中在 {X}-{Y} 万字
- 书名特征:{命名模式总结}
- 标签热词:{高频标签词}
### 值得关注的方向
1. {方向 + 为什么值得关注 + 可行性评估}
2. {方向 + 为什么值得关注 + 可行性评估}
3. {方向 + 为什么值得关注 + 可行性评估}
### 一句话
{犀利的总结}
Phase 5: Topic decision
Follow the four topic-selection steps, feasibility rules, and template in references/topic-decision.md to complete 选题决策.md。
Continue from the file: first read the "List-scan conclusions" in {outdir}/选题决策.md and 扫榜聚合.md, not relying on conversation memory; if neither exists, go back to Phase 2.
If the information is insufficient, ask the user to fill in the project conditions:"What are the target platform, existing materials, genres you are good at/writing constraints, and planned length?"
According to topic-decision.md four-step topic selection produces 2-3 recommended topics (reason it can go viral → market validation → differentiated positioning → feasibility + failure risk + validation actions), written intothis scan output directory {outdir}/选题决策.md in the "Recommended Topics" section, keeping the "Scan Conclusions".
Hard rules:
- Feasibility ceiling: when the ranking is marked
[数据稀疏], the aggregate table is marked "few", or same-direction samples <15 (small platforms <10), give at most "Medium" and state that validation comes first; built-in knowledge likewise gives at most "Medium". When materials, platform, or other conditions do not match, still give "Low". - Topics must be supported by project materials, state feasibility and failure risk clearly, and fit the target platform's tone.
Platform Characteristics Quick Reference
| Platform | Tone | Core Metrics | Primary Readers | Suitable Genres |
|---|---|---|---|---|
| Qidian Chinese Network | Mainly male-oriented, hardcore wish-fulfillment | Monthly ticket/bestseller ranking position | Audience for leveling up, growth, and long-term serialized reading | Fantasy, urban, sci-fi, game |
| Fanqie Novel | Lower-tier market, free reading | Reads, reading ranking position | General readers | High-concept, fast-paced, strongly satisfying |
| Jinjiang Literature City | Mainly female-oriented | Favorites, nutrient solution, points | Romance, danmei, and derivative audience | Romance, danmei, derivative |
| Qimao Novel | Lower-tier market, free reading | Popularity, hot ranking position | General readers | Fast-paced wish-fulfillment |
| Ciwimao | ACG, light novels | Values of various rankings | ACG audience | Fanfiction, ACG, light novels |
Process Handoff
Pipeline: long-form Position: scan rankings (step 1/3)
| Timing | Jump to | Command |
|---|---|---|
| Found a direction | story-long-analyze | /story-long-analyze |
| Start writing directly | story-long-write | /story-long-write |
| More suitable for short-form | story-short-scan | /story-short-scan |
Reference Materials
Load the following files as needed:
| File | When to Load |
|---|---|
| references/topic-decision.md | "Topic Decision": four-step topic selection + feasibility judgment + Topic Decision.md template |
| references/reader-profiling.md | When you need to analyze the target reader profile |
| references/genre-trends.md | When checking genre trend candidates, entry constraints, and sample validation rules |
| references/publishing-guide.md | Platform adaptation + recommendation mechanism validation + data metrics + synopsis design |
| Five platform references (links see "Script Collection Mode" step 1) | "Determine Data Source": read only the one for the selected platform—ranking URLs, commands, fields, troubleshooting, platform analysis dimensions |
| scripts/aggregate-rank.js | Aggregate the raw rankings in the output directory into a short table (--out saved to disk,--sample extract raw entries,--json machine-readable), the main session reads the aggregate table first |
| scripts/cdp-utils.js and each platform's collection scripts | CDP common tools and each platform's collection scripts ({平台}-rank-scraper.js), commands see the selected platform reference |