Back to Details

xiaohongshu-content-benchmark

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

XHS Content Benchmark

XHS内容基准分析

Follow shared release-shell rules in:
  • postplus-shared
    release-shell rules
Legacy alias: 
xhs-content-benchmark
.
Use this skill when the user wants to:
  • benchmark competitor or inspiration accounts on Xiaohongshu
  • identify repeated Xiaohongshu title hooks and cover patterns
  • shortlist strong note examples before adapting them into local XHS cards
  • compare recent post output from one or more known Xiaohongshu profiles
Read these references before implementation:
  • skills/20-research/xhs-content-benchmark/references/actor-selection.md
  • skills/20-research/xhs-content-benchmark/references/brief-schema.md
  • skills/20-research/xhs-content-benchmark/references/normalized-schema.md
遵循以下共享的release-shell规则：
  • postplus-shared
    release-shell规则
旧别名：
xhs-content-benchmark
.
当用户有以下需求时使用此技能：
  • 对小红书竞品或灵感账号进行基准分析
  • 识别小红书重复出现的标题钩子和封面样式
  • 在将优质笔记改编为本地XHS卡片前筛选出优秀示例
  • 对比一个或多个已知小红书账号的近期发布内容
执行前请阅读以下参考文档：
  • skills/20-research/xhs-content-benchmark/references/actor-selection.md
  • skills/20-research/xhs-content-benchmark/references/brief-schema.md
  • skills/20-research/xhs-content-benchmark/references/normalized-schema.md

Default posture

默认操作方式

Start from account-based benchmarking, not broad keyword search.
Default starting surface:
  • known profile URLs -> recent post benchmark pool
Do not use this as the default first surface:
  • keyword or topic search as the first collection surface
Do not silently switch between surfaces. If the user asks for keyword search, use the keyword-search route explicitly and fail visibly if it returns no items.
从基于账号的基准分析开始，而非宽泛的关键词搜索。
默认起始数据源：
  • 已知账号URL -> 近期帖子基准分析池
不要将以下方式作为默认首选数据源：
  • 将关键词或话题搜索作为首个内容收集渠道
不要在不同数据源间静默切换。如果用户要求关键词搜索，请明确使用关键词搜索路径，若返回无结果则需清晰提示失败。

Validated actor path

已验证的执行路径

Default actor for this skill:
  • rednote-xiaohongshu-user-posts-scraper
Use it when the user provides:
  • one or more Xiaohongshu 
    profileUrls
  • one or more Xiaohongshu 
    profileIds
This is the current default because it returns the post-level fields this benchmark flow needs:
  • note URL
  • note id
  • title
  • post type
  • like count
  • author metadata
  • cover image metadata
此技能的默认执行工具：
  • rednote-xiaohongshu-user-posts-scraper
当用户提供以下信息时使用该工具：
  • 一个或多个小红书
    profileUrls
  • 一个或多个小红书
    profileIds
这是当前默认工具，因为它能返回基准分析流程所需的帖子级字段：
  • 笔记URL
  • 笔记ID
  • 标题
  • 帖子类型
  • 点赞数
  • 作者元数据
  • 封面图片元数据

Experimental route

实验性路径

Keyword route:
  • rednote-xiaohongshu-search-scraper
Use it only when the user explicitly wants keyword or topic benchmarking.
Current operational rule:
  • require 
    maxItems >= 100
  • fail if the actor returns zero items
Do not pretend keyword benchmarking succeeded when the actor returned an empty dataset.
关键词路径：
  • rednote-xiaohongshu-search-scraper
仅当用户明确要求关键词或话题基准分析时使用此路径。
当前操作规则：
  • 要求
    maxItems >= 100
  • 若工具返回空结果则提示失败
当工具返回空数据集时，不得伪装关键词基准分析成功。

Not part of this skill

不属于此技能的范围

  • comment mining
  • audience language extraction
  • media downloading
  • profile health snapshots
  • publishing or scheduling notes
Route those later to dedicated XHS skills.
  • 评论挖掘
  • 受众语言提取
  • 媒体下载
  • 账号健康快照
  • 笔记发布或排期
上述需求请转至专门的XHS技能处理。

Recommended workflow

推荐工作流程

  1. classify the request:
    • account benchmark
    • keyword benchmark
  2. compile a small benchmark brief into actor input
  3. run the chosen actor through the shared collection runner
  4. normalize into one local post dataset
  5. rank by like signal plus theme relevance
  6. summarize:
    • recurring hooks
    • title pattern families
    • cover aspect patterns
    • content-type split
    • strongest benchmark examples
  1. 对请求进行分类：
    • 账号基准分析
    • 关键词基准分析
  2. 将小型基准分析简报整理为工具输入参数
  3. 通过共享收集运行器执行所选工具
  4. 归一化为本地帖子数据集
  5. 按点赞量和主题相关性排序
  6. 总结输出：
    • 重复出现的标题钩子
    • 标题模式分类
    • 封面比例样式
    • 内容类型划分
    • 最优基准分析示例

Cost discipline

成本控制规则

Validated account route:
  • start with 8-15 recent posts per profile
  • benchmark 1-3 profiles at a time
Experimental keyword route:
  • the actor currently enforces a much higher minimum item count
  • do not hide that constraint
  • keep keyword batches narrow and specific
已验证的账号路径：
  • 每个账号先取8-15条近期帖子
  • 每次基准分析1-3个账号
实验性关键词路径：
  • 当前工具强制要求更高的最小条目数
  • 不得隐瞒该限制
  • 关键词批次需保持精准且范围狭窄

Release-Shell Execution Contract

Release-Shell执行约定

  • keep benchmark briefs, actor inputs, raw datasets, normalized datasets, rankings, and analysis outputs under 
    <work-folder>/.postplus/xiaohongshu-content-benchmark/
  • keep only final user-facing benchmark summaries or shortlists outside 
    .postplus/
  • start with a bounded first pass: 
    • 1-3
      profiles for the validated account route
    • narrow keyword batches only when the user explicitly requests search
  • if hosted capability is unavailable, unauthorized, or returns a stable network error, stop immediately instead of switching to ad hoc shell glue
  • 将基准分析简报、工具输入、原始数据集、归一化数据集、排序结果和分析输出存储在 
    <work-folder>/.postplus/xiaohongshu-content-benchmark/
    目录下
  • 仅将最终面向用户的基准分析总结或筛选列表存放在
    .postplus/
    目录外
  • 首次执行需限定范围：
    • 已验证账号路径最多分析
      1-3
      个账号
    • 仅当用户明确要求搜索时才使用范围狭窄的关键词批次
  • 若托管功能不可用、未授权或返回稳定网络错误，需立即停止操作，不得改用临时脚本

Main scripts

主要脚本

  • scripts/build_xhs_benchmark_actor_input.mjs
  • scripts/normalize_xhs_benchmark_dataset.mjs
  • scripts/rank_xhs_benchmark_posts.mjs
  • scripts/analyze_xhs_benchmark_dataset.mjs
Use the shared collection runner for actor calls:
  • ${CLAUDE_SKILL_DIR}/_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs
  • scripts/build_xhs_benchmark_actor_input.mjs
  • scripts/normalize_xhs_benchmark_dataset.mjs
  • scripts/rank_xhs_benchmark_posts.mjs
  • scripts/analyze_xhs_benchmark_dataset.mjs
使用共享收集运行器调用工具：
  • ${CLAUDE_SKILL_DIR}/_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs

Minimal workflow

极简工作流程

Account benchmark

账号基准分析

Build actor input from a brief:
bash
node ${CLAUDE_SKILL_DIR}/scripts/build_xhs_benchmark_actor_input.mjs \
  --brief <work-folder>/.postplus/xhs-benchmark-brief.json \
  --output <work-folder>/.postplus/xhs-benchmark-actor-input.json
Run the validated actor:
bash
node ${CLAUDE_SKILL_DIR}/_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs \
  --collection-key xhs-account-posts \
  --input <work-folder>/.postplus/xhs-benchmark-actor-input.json \
  --output <work-folder>/.postplus/xhs-benchmark-raw.json
Normalize, rank, analyze:
bash
node ${CLAUDE_SKILL_DIR}/scripts/normalize_xhs_benchmark_dataset.mjs \
  --input <work-folder>/.postplus/xhs-benchmark-raw.json \
  --output <work-folder>/.postplus/xhs-benchmark-normalized.json

node ${CLAUDE_SKILL_DIR}/scripts/rank_xhs_benchmark_posts.mjs \
  --input <work-folder>/.postplus/xhs-benchmark-normalized.json \
  --theme "职场,打工人,办公室" \
  --output <work-folder>/.postplus/xhs-benchmark-ranking.json

node ${CLAUDE_SKILL_DIR}/scripts/analyze_xhs_benchmark_dataset.mjs \
  --input <work-folder>/.postplus/xhs-benchmark-normalized.json \
  --output <work-folder>/.postplus/xhs-benchmark-analysis.json
通过简报构建工具输入：
bash
node ${CLAUDE_SKILL_DIR}/scripts/build_xhs_benchmark_actor_input.mjs \
  --brief <work-folder>/.postplus/xhs-benchmark-brief.json \
  --output <work-folder>/.postplus/xhs-benchmark-actor-input.json
执行已验证工具：
bash
node ${CLAUDE_SKILL_DIR}/_postplus_shared/00-core/shared-collection/scripts/collection_actor_run.mjs \
  --collection-key xhs-account-posts \
  --input <work-folder>/.postplus/xhs-benchmark-actor-input.json \
  --output <work-folder>/.postplus/xhs-benchmark-raw.json
归一化、排序、分析：
bash
node ${CLAUDE_SKILL_DIR}/scripts/normalize_xhs_benchmark_dataset.mjs \
  --input <work-folder>/.postplus/xhs-benchmark-raw.json \
  --output <work-folder>/.postplus/xhs-benchmark-normalized.json

node ${CLAUDE_SKILL_DIR}/scripts/rank_xhs_benchmark_posts.mjs \
  --input <work-folder>/.postplus/xhs-benchmark-normalized.json \
  --theme "职场,打工人,办公室" \
  --output <work-folder>/.postplus/xhs-benchmark-ranking.json

node ${CLAUDE_SKILL_DIR}/scripts/analyze_xhs_benchmark_dataset.mjs \
  --input <work-folder>/.postplus/xhs-benchmark-normalized.json \
  --output <work-folder>/.postplus/xhs-benchmark-analysis.json

Keyword benchmark

关键词基准分析

Use the same build script, but provide 
keywords
in the brief.
Operational constraint:
  • keep 
    limit >= 100
  • expect a visible failure if the actor returns zero items
使用相同的构建脚本，但需在简报中提供
keywords
参数。
操作约束：
  • 保持
    limit >= 100
  • 若工具返回空结果，需清晰提示失败

Good output

优质输出内容

Return:
  • benchmark shortlist
  • top note URLs
  • recurring title hooks
  • title pattern breakdown
  • cover-shape breakdown
  • observed content-type split
  • data-quality warnings when fields are missing
  • suggested card angles that can later feed 
    skills/40-creative/xiaohongshu-card-notes
返回：
  • 基准分析筛选列表
  • 顶级笔记URL
  • 重复出现的标题钩子
  • 标题模式细分
  • 封面样式细分
  • 已识别的内容类型划分
  • 字段缺失时的数据质量警告
  • 可用于后续
    skills/40-creative/xiaohongshu-card-notes
    的卡片创作建议角度

Failure posture

失败处理方式

  • fail if the request mixes profile benchmarking and keyword benchmarking without saying which should be primary
  • fail if no 
    profileUrls
    , 
    profileIds
    , or 
    keywords
    are provided
  • fail if the keyword-search actor returns zero items
  • fail if the normalized dataset has no note URLs
  • keep raw actor output for debugging
  • 若请求同时包含账号基准分析和关键词基准分析但未明确主次，提示失败
  • 若未提供
    profileUrls
    、
    profileIds
    或
    keywords
    ，提示失败
  • 若关键词搜索工具返回空结果，提示失败
  • 若归一化数据集无笔记URL，提示失败
  • 保留工具原始输出用于调试