SuperAPI AI Directory
Partner中文AI newsFavorites
Skillskuaishou-comment

kuaishou-comment

General & other

View repository

Updated 10-16 01:48

About this skill

Kuaishou Comment Analysis

📝 Overview

Enter a Kuaishou work link to get first-level comment data, 20 per page, supporting cursor-based pagination. Display all comments on the current page and AI summary analysis in the conversation, while generating an interactive HTML report containing the current page data and summary.

✨ Features

Feature Module Capability Description Core Value
Comment Retrieval Paste a work link to get comments Fetch work comment section data in real time
Paginated Browsing 20 per page, cursor-based pagination Browse large numbers of comments page by page
AI Summary Four-dimensional sentiment analysis (positive/negative/demand/competitor) Quickly understand the full picture of comment public opinion
Pinned Marking Special marking for pinned comments Identify important comments at a glance
HTML Report Dark theme interactive report Save/share analysis results offline
Merged Report Multi-page merge, auto-collapse beyond 3 pages Export all comment analysis at once

🔑 Authentication

Go to Red Fox Hub to get an API Key, and configure it in the following ways:

# 方式一：配置文件（如 OpenClaw 的 ~/.openclaw/openclaw.json）
{ "env": { "REDFOX_API_KEY": "ak_xxxx..." } }

# 方式二：终端环境变量
export REDFOX_API_KEY="ak_xxxx..."

🔄 Workflow

Step 1: Understand user intent, extract opusId

⚠️ Core rule: The user provides a work link, and the Agent extracts the opusId from the link.

  • The user provides a work link (e.g., https://www.kuaishou.com/short-video/3x4ibfzs5e68yxu), and the Agent extracts the opusId from the link
  • The user may also directly provide an opusId (e.g., 3x4ibfzs5e68yxu), in which case use it directly
  • If the user does not provide a work link, proactively ask: "Please provide a Kuaishou work link"
  • If the user queried data for a certain work in the previous round of conversation, and the current input is ambiguous (such as "next page" or "comment analysis"), use the opusId from the previous round

Step 2: Call the comment retrieval script (automatically generates HTML)

⚠️ Each call requests only one page of data (one API request); do not initiate multiple calls to fetch multiple pages without authorization.

The script automatically generates an HTML report on each call(contains placeholders, analysis data not yet filled in). Terminal display and HTML are based onthe same API call, ensuring data consistency.

python3 ~/.agents/skills/kuaishou-comment/scripts/kuaishou_comment_search.py "<opusId>" [--cursor <cursor>] [--page <page>]

Parameter description:

Parameter Description Default Value
opusId Kuaishou work opusId (required) —
--cursor Cursor; pass an empty string for the first page, and pass the nextCursor returned by the previous page for subsequent pages ""
--page Current page number (only used for the HTML file name and output marker) 1
--output-dir HTML output directory ~/Downloads/QoderReports

For the JSON fields returned by the script, see the comment field table below.

Note: The HTML report currently contains only the raw comment data; the analysis placeholders will be filled in after user confirmation.

Step 3: Display comment data + AI summary in the conversation

⚠️ Each operation calls the API only once to fetch one page of data.

A1. Inform the query scope

📊 Work「{opusId}」retrieved a total of Ncomments (page {page}); the detailed data is as follows:

A2. Render the comment table (display all comments on the current page)

Formatting rules:

  • The commenter column displays only the nickname; no avatar is shown and no hyperlink is generated
  • Comment content exceeding 40 characters is truncated with ...
  • Time is formatted as MM-DD HH:MM
  • Pinned comments get a 📌 marker after the nickname (the entire row is bolded)
  • like_count / reply_count numbers ≥ 10000 use the x.xw format
  • Display everything on each page; no truncation allowed

A3. AI comment summary (⚠️ must be output for every query, including page turns)

First page: Based on the comments on the current page, four-dimensional sentiment analysis.

After page turn: Based on comprehensive analysis of the cumulative data of all pages retrieved so far, marking the cumulative total count.

Before analysis, one must understand Kuaishou community culture(the "laotie" culture, emoji semantics, and unique expressions such as "aoligei" and "double-tap").

📈 Comment summary analysis (based on {total} comments)

✅ Positive reviews ({positive_ratio}%)

  • {key points, citing representative comment keywords}

⚠️ Negative reviews ({negative_ratio}%)

  • {key points}

💡 User needs ({demand_ratio}%)

  • {key points}

🔍 Competitor comparison public sentiment ({competitor_ratio}%)

  • {output only when competitors are mentioned; otherwise output "No competitors mentioned"}

Analysis requirements:

  • Each key point cites representative comment keywords/phrases
  • The percentage is the proportion of that type of comment; the sum of the four categories may exceed 100%
  • Must understand Kuaishou community culture

A4. Page-turn prompt (⚠️ immediately after A3)

  • If has_next is true:📄 当前第 {page} 页（共 {total_fetched} 条）。回复「下一页」继续查看。
  • If has_next is false:📄 当前第 {page} 页，已无更多数据。

⚠️ A1~A4 must be completed consecutively in the same round of output; no step may be omitted.

After A4 output is complete,ask the user whether an HTML visualization report needs to be generated：

📊 Do you need to generate an HTML visualization report?

If the user confirms, choose the method based on the number of pages browsed:

Only 1 page → single-page HTML (backfill_html.py)

The HTML file was already generated in Step 2; directly backfill the analysis data and open it:

python3 ~/.agents/skills/kuaishou-comment/scripts/backfill_html.py "<html_path>" --analysis-json '<分析JSON>'
open "<html_path>"

2 pages or more → consolidated report (consolidate_report.py)

Collect the JSON data of all browsed pages and generate one containingall comments and cumulative analysiscomplete HTML. Display in sections by page number,automatically collapse when there are more than 3 pages, and prompt the user to click to expand.

python3 ~/.agents/skills/kuaishou-comment/scripts/consolidate_report.py "<opusId>" \
  --pages-json '[{"page":1,"total":31,"comments":[...]},{"page":2,"total":20,"comments":[...]}]' \
  --analysis-json '<累计分析JSON>'
open "<html_path>"
Parameter Description
opusId Kuaishou work opusId (positional parameter)
--pages-json Multi-page data JSON array, each item containing page, total, comments
--analysis-json Cumulative analysis JSON (fields same as backfill)

pages data is collected from the output of each script call; cumulative analysis is based on the comprehensive analysis after A3 page turning.

Analysis JSON field description:

Field Type Description
positive_ratio string Positive review percentage (pure number, e.g. "62", the template automatically appends %)
negative_ratio string Negative review percentage
demand_ratio string User demand percentage
competitor_ratio string Competitor comparison percentage
positive_summary string Positive review summary (HTML ul format)
negative_summary string Negative review summary
demand_summary string User demand summary
competitor_summary string Competitor comparison summary

Step 4: Page turning handling (when the user replies "next page", etc.)

  1. Reuse opusId
  2. Use the nextCursor from the previous page
  3. Call the script again:python3 ~/.agents/skills/kuaishou-comment/scripts/kuaishou_comment_search.py "<opusId>" --cursor "<nextCursor>" --page <page>
  4. Fully execute A1~A4, with A3 based on cumulative comprehensive analysis
  5. After A4, ask whether HTML is needed

Page turning rules: has_next true → use nextCursor to turn the page; false → no more. Pass cursor through as-is; do not modify it.

Step 5: Error handling

Error type Handling method
No API Key Prompt to configure REDFOX_API_KEY
Invalid work link Prompt "No reviews found for this work, please check whether the work link is correct"
API returned an error Display the error code and error message
0 reviews retrieved Prompt "This work currently has no reviews"
Network request timeout Prompt "Network request timed out, please try again later"

📋 Dependencies

The script uses urllib (Python standard library), no additional dependencies need to be installed.

🎯 Usage examples

Example 1: View reviews for a specified work

用户：查看快手作品 3x4ibfzs5e68yxu 的评论
助手：调用脚本 → 展示当前页 + AI总结 → 询问HTML → backfill → open

Example 2: Query via link

用户：https://www.kuaishou.com/short-video/3x4ibfzs5e68yxu 这个作品的评论怎么样
助手：提取 opusId → 调用脚本 → 展示分析 → 询问HTML

Example 3: Merge and export after page turning

用户：下一页 → 下一页 → 生成html
助手：收集3页JSON → consolidate_report.py → 1个HTML含全部评论+折叠 → open