Task Objective
- This Skill is used to: help Xiaohongshu creators find benchmark accounts
- Capabilities include: same-tier benchmark matching (similar follower count), higher-tier benchmark recommendations (3-5x follower count), account data analysis
- Trigger conditions: when users express needs such as "find benchmark accounts", "recommend Xiaohongshu accounts", "KOL placement", "account launch reference", etc.
Prerequisites
- Dependency description: Python standard library (json, argparse, socket, ssl)
- API description: call the Xiaohongshu account recommendation API (use native socket+ssl to manually send HTTPS requests, without sending SNI)
Operation Steps
Step 1: Receive user input and parse parameters
Input method A: Xiaohongshu account ID
- User input example: "My ID is 27493135897"
- Parameter: redId
Input method B: track + follower count + account level (optional)
- User input example: "cooking, 3000 followers, amateur"
- Parameter mapping:
- track: track type (automatically mapped, e.g. "cooking"→"Delicious Cuisine")
- maxFans/minFans: follower count processing rules are shown below
- level: account level (automatically mapped, e.g. "newbie"→"amateur")
Follower count processing rules (important):
- Single follower count (e.g. "3000 followers"): maxFans=3000, minFans=0(do not pass an empty string)
- Follower range (e.g. "1000-3000 followers"): maxFans=3000, minFans=1000
- No follower count specified: pass empty strings for both maxFans and minFans
Step 2: Call the script to query benchmark accounts
# 按账号ID查询
python scripts/xiaohongshu_account_recommender.py --red_id "27493135897"
# 按赛道+粉丝数+等级查询
python scripts/xiaohongshu_account_recommender.py --track "美味佳肴" --min_fans 0 --max_fans 3000 --level "素人"
API endpoint: POST https://onetotenvip.com/story/xhsUser/querySimilarAccounts
Step 3: Output results according to the standard template
Regardless of the query method, the output format is completely identical, strictly follow references/output_template.md output into the conversation:
- Opening statement (only show groups with data; when there is no data, show "No matches found yet")
- Same-tier benchmark table (show only if there is data)
- Higher-tier benchmark table (show only if there is data)
- Analysis summary (show only if there is data)
- Subscription service
Recommendation reason generation rules: comprehensively analyze content style, update frequency, interaction level, and viral post cases, and describe in colloquial language.
Step 4: Generate and display an HTML report (mandatory)
After the account recommendation output is completed, the script will automatically execute the following process:
- Store all data in a temporary JSON file:
/workspace/projects/xiaohongshu-account-recommender/account_data.json - Read the JSON file data to generate an HTML file:
/workspace/projects/xiaohongshu-account-recommender/account_recommend.html - The script outputs a JSON marker:
{"status": "success", "html_path": "..."}
AI must execute: read the script outputhtml_pathpath corresponding to the HTML file, and directly display that HTML file at the end of the conversation for the user to preview.
Standard track categories (25)
All | Travel and Commuting | Leisure and Hobbies | Film and Entertainment | Digital Technology | Healthcare | General Miscellaneous | Astrology and Relationships | Fashion and Outfits | Weddings and Marriage | Photography and Recording | Learning and Education | Makeup and Beauty | Home and Renovation | Travel and Vacation | Parenting and Childcare | Personal Care | Delicious Food | Career Development | Pet World | Trendy Shoes and Bags | Daily Life | Scientific Exploration | News and Information | Physical Exercise
Standard Account Levels (7)
Celebrity | Brand | Enterprise | Top KOL | Mid-tier KOL | Tail KOL | Regular Person
Track Mapping Example
| User Input | Match Result |
|---|---|
| Beauty, Makeup, Skincare | Makeup and Beauty |
| Outfits, Fashion | Fashion and Outfits |
| Food, Cooking, Store Visits | Delicious Food |
| Travel, Tourism | Travel and Vacation |
| Home, Renovation, Makeover | Home and Renovation |
| Fitness, Sports, Weight Loss | Physical Exercise |
| Mother and Baby, Parenting | Parenting and Childcare |
| Pets, Cats, Dogs | Pet World |
| Digital, Mobile Phones, Technology, Internet, AI, Artificial Intelligence, Programming | Digital Technology |
| Education, Learning | Learning and Education |
| Relationships, Romance | Astrology and Relationships |
| Workplace, Work | Career Development |
Account Level Mapping Example
| User Input | Match Result |
|---|---|
| Celebrity, Artist, Idol, idol | Celebrity |
| Brand, Label | Brand |
| Enterprise, Company, Merchant, Official Account | Enterprise |
| Top KOL, Top, Top Influencer, Big V, Big Account, Million Followers | Top KOL |
| Mid-tier KOL, Mid-tier, Mid-tier Influencer, Medium Account, 100k Followers | Mid-tier KOL |
| Tail KOL, Tail, Tail Influencer, Small Account, 10k Followers | Tail KOL |
| Regular Person, Ordinary Person, Novice, Newbie, Newcomer, Newbie Blogger, Personal Account | Regular Person |
Usage Examples
Example 1: Query by Account ID
- User Input: "My ID is 27493135897"
- Script Call:
python scripts/xiaohongshu_account_recommender.py --red_id "27493135897" - Output Format: Output according to the standard template
Example 2: Query by Track + Single Follower Count + Level
- User Input: "Cooking, 3000 followers, Regular Person"
- Parameter Mapping: track=Delicious Food, minFans=0, maxFans=3000, level=Regular Person
- Script Call:
python scripts/xiaohongshu_account_recommender.py --track "美味佳肴" --min_fans 0 --max_fans 3000 --level "素人" - Output Format: Output according to the standard template (exactly the same as Example 1)
Example 3: Query by Track + Follower Range + Level
- User Input: "Cooking, 1000-3000 followers, Regular Person"
- Parameter Mapping: track=Delicious Food, minFans=1000, maxFans=3000, level=Regular Person
- Script Call:
python scripts/xiaohongshu_account_recommender.py --track "美味佳肴" --min_fans 1000 --max_fans 3000 --level "素人" - Output Format: Output according to the standard template (exactly the same as Example 1)
Resource Index
- Script: See scripts/xiaohongshu_account_recommender.py(Call the API to query benchmark accounts, supports track and account-level mapping, automatically generates an HTML report)
- Reference: see references/output_template.md(Output format specification, template definition, numeric formatting rules)
- Asset: see references/account_template.html(HTML report template, consistent with the output template format)
Notes
- Follower count handling: single value → maxFans=that value, minFans=0; range → pass parameters by range
- Regardless of the query method used, the output format is completely identical, strictly follow
references/output_template.mdfor execution - All data comes from API interfaces; online searching is strictly prohibited