SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillswechat-article-search

wechat-article-search

Writing & research

View repository

Updated 10-16 04:57

About this skill

WeChat Official Account Article Search Instructions

Applicable Scenarios

  • The user says "help me search for Official Account articles/recent articles with a certain keyword"
  • Need to quickly obtain: title, summary, publish time, Official Account name, accessible link

Workflow

Step 1: Locate the script and dependencies

First set SKILL_DIR to the directory containing the SKILL.md read this time (resolve symlinks to the real directory first). Dependencies are managed within that directory; if already installed, use them directly. Only when dependencies are missing, install them according to the repository conventions below, do not install globally, and do not install in the user's project root directory.

Step 2: Use existing input

Extract keywords from the user request; if no quantity is specified, use the default of 10 items. If keywords and scope already exist, execute directly; only ask when the search topic is missing. Output files are created only when the user requests saving.

Step 3: Execute the search command

1. Execute the regular search command

node "$SKILL_DIR/scripts/search_wechat.js" "关键词"

Special process (optional)

  1. Execute the search command including a quantity limit
node "$SKILL_DIR/scripts/search_wechat.js" "关键词"  -n 15
  1. If the user needs to save the results to a file, execute the command
node "$SKILL_DIR/scripts/search_wechat.js" "关键词" -n 20 -o result.json
  1. If you want to obtain the real link of the WeChat article domain, execute the following command
node "$SKILL_DIR/scripts/search_wechat.js" "关键词" -n 5 -r

Parameter Description

  • query: Search keyword (required)
  • -n, --num: Number of results to return (default 10, maximum 50)
  • -o, --output: Output JSON file path (optional)
  • -r, --resolve-url: Attempt to resolve the intermediate link into the real WeChat article link (will make an additional request for each result)

Output Fields (Article Object)

Article title, article address, article summary, publish time, source Official Account name

Common Issue Handling

  • Empty results: try changing keywords, using fewer special characters, or retrying later
  • Failed to resolve the real URL: this is normal (anti-crawling restrictions); you can prompt the user to open the intermediate link in a browser

Notes

  • This tool is for learning and research purposes only. Do not use it for commercial purposes or large-scale crawling.
  • When using this tool, please comply with the terms of use and regulations of the relevant websites.
  • Excessive use may result in IP bans, please use with caution.

Vault Adaptation Notes (exclusive to deepsight_vault, this section is not in the original version)

The following content is automatically appended by Dapeng's vault to integrate this skill into the Obsidian repository ecosystem. See the original instructions above.

Dependency Installation (Conventions of This Repository)

This skill is self-contained with dependencies within the repository,do notinstall globally. On first use or after cloning the repository, execute once in this skill directory:

cd "$SKILL_DIR"
npm install

cheerio will be installed into this directory's node_modules/, which has been excluded by .gitignore . The script can use Node's upward lookup mechanism to require('cheerio'), without needing to configure NODE_PATH。

Single Source of Truth and Synchronization

  • Source of truth: this Portfolio's skills/wechat-article-search/。
  • The installation entry point can be a symlink within the host Skill directory; the actual directory resolved from the path read this time shall prevail, and it is not assumed that all hosts share the same sync script.
  • Modify the skills/wechat-article-search/ real source of this Portfolio; resolve the symlink first, and do not treat the installation entry point as another copy of the source code.

Connection with wxmp-article-harvester integration with

The responsibilities of the two skills are complementary and can be used in sequence:

Stage skill Input Output
Keyword discovery (this skill) wechat-article-search Keywords, e.g. "AI Agent" Article list JSON (title/link/summary/source/time)
Fetch article body by account/URL wxmp-article-harvester Official account name or mp.weixin.qq.com/s/... link Structured index + Markdown article body

Typical flow: First use this skill to discover a batch of article URLs and source official accounts by keyword → then use harvester to fetch and export the article bodies for accounts or specific URLs of interest to ~/.dapeng/wxmp-harvester/exports/。

Output destination suggestions

  • Temporary retrieval: directly view the stdout JSON without writing to disk.
  • When archiving is needed: use -o to write to 00_收件箱/Clippings/ or the corresponding project's source directory, with the file name containing the keyword and date, e.g. wechat-search-AI培训-2026-07-06.json。
  • Do not scatter search results into the repository root directory.

Anti-scraping and disclaimer (important)

  • The data source is Sogou WeChat Search (weixin.sogou.com), which is affected by anti-scraping policies;occasionally returning empty results or being blocked by antispider is normal, and it is sufficient to change keywords or retry later.
  • -r Resolving real links has a low success rate when anti-scraping is tightened; on failure, the Sogou intermediate link will be retained and marked url_resolved: false。
  • For learning, research, and personal material organization only; do not call it at high frequency and do not perform large-scale commercial scraping.