xhs-comment-insight

Analyzes Xiaohongshu note comments to generate one new content topic as structured JSON.

1.6k|214|Updated Jun 4, 2025
One-click install
npx skills add https://github.com/Jamailar/RedBox --skill xhs-comment-insight
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: xhs-comment-insight
Source: https://github.com/Jamailar/RedBox/tree/main/desktop/builtin-skills/xhs-comment-insight
Command: npx skills add https://github.com/Jamailar/RedBox --skill xhs-comment-insight

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Content creators on Xiaohongshu often struggle to find their next topic. Comment sections contain real audience questions, objections, and unmet needs, but manually mining them for a focused, writable topic is slow and inconsistent.

Core Features & Use Cases

  • Comment Signal Classification: Categorizes comments into follow-up questions, rebuttals, supplements, emotions, action intents, and author interactions to surface what readers still want to know.
  • Single-Topic Generation: Distills one note's comment section into exactly one narrow, writable topic with a title under 20 characters, target reader, core tension, angle, and material source.
  • Strict JSON Output: Returns a fixed schema with content_direction, thinking_process, topic, and direction_frame fields, ready for downstream topic-center pipelines.
  • Use Case: A creator picks a viral note about budget travel; the skill reads its comments, finds many users asking how to book cheap train tickets, and outputs a new topic targeting first-time budget travelers.

Quick Start

Provide one Xiaohongshu note with its comment excerpts and ask the AI to analyze the comment section and produce one potential topic.

Frequently Asked Questions about xhs-comment-insight

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find new Xiaohongshu topics from comments?

Provide one Xiaohongshu note together with its comment excerpts, and the skill classifies comment signals such as follow-up questions, rebuttals, and action requests, then outputs one narrow topic with a title, target reader, and angle in strict JSON format.

What input does comment insight analysis require?

It requires exactly one note and its comment excerpts for the current round. If no note or comments are provided in chat, the skill asks you to supply comment material or select a note with comments before proceeding.

Does the analysis use my account profile or knowledge base?

No. The skill explicitly excludes long-term memory, account positioning, user profiles, and other knowledge base content. It only uses the current note, its comments, and optionally the same note's comments.json, meta.json, or body file.

What output format does the comment analysis produce?

It outputs a single JSON object with content_direction, thinking_process, topic, and direction_frame fields. The topic title must be under 20 characters and topic.connections must be [1], with no Markdown or code blocks.

When should I not use comment section mining for topics?

Avoid it when you need cross-note synthesis, comment volume statistics, or summaries of the original note. The skill is designed only to find one small, writable angle from a single note's comment signals, not broad research questions.