wechat-article-source

name	wechat-article-source
description	Read, extract, summarize, and source-check WeChat Official Account articles from mp.weixin.qq.com links. Use when a user asks to inspect a WeChat article, summarize key points, recover article metadata/body text, troubleshoot blocked WeChat article fetching, or find and verify related original information sources.

WeChat Article Source

Objective

Turn a WeChat Official Account article URL into an evidence-backed brief:

• article title, account name, publication time, source URL, and canonical identifiers when available
• readable article body and in-article links
• concise summary of the article's main claims
• related source map, separating verified primary sources, third-party reports, and unverifiable author-only claims
• clear note about access limitations, anti-bot blocks, login walls, or weak evidence

Workflow

1. Start from the real URL the user gave. Do not infer the article title from search results unless direct article access fails.
2. Try browser-visible access first when an in-app browser or logged-in Chrome is already available.
3. If browser access is unavailable, blocked, or too slow, fetch the HTML directly.
4. If local network fails because proxy variables point at a dead local proxy, retry with proxy variables removed.
5. If sandbox DNS or network restrictions block fetching, request approval to rerun the same fetch outside the sandbox.
6. Extract metadata and正文 from the returned HTML.
7. Search for sources using article title, product/person names, unusual phrases, URLs, and claims from the body.
8. Prefer primary sources: official sites, official docs, public APIs, repositories, regulatory filings, papers, or original posts. Use third-party articles as secondary context.
9. In the final answer, distinguish:
   • what came from the WeChat article itself
   • what was verified against external sources
   • what remains author-only or not publicly verifiable

Fetching Pattern

Use a normal browser User-Agent. Many WeChat article pages embed the full content in the initial HTML.

If curl fails with a local proxy error such as Failed to connect to 127.0.0.1 port 10808, remove proxy variables:

env -u ALL_PROXY -u all_proxy -u HTTP_PROXY -u HTTPS_PROXY -u http_proxy -u https_proxy \
  curl -L -s -A 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125 Safari/537.36' \
  'https://mp.weixin.qq.com/s/ARTICLE_ID'

If the command then fails with DNS or sandbox-network errors, rerun with the platform approval mechanism instead of inventing alternate URLs.

Extraction Script

Use scripts/extract_wechat_article.py for deterministic extraction from either a URL or a saved HTML file.

Examples:

python3 wechat-article-source/scripts/extract_wechat_article.py \
  --url 'https://mp.weixin.qq.com/s/ARTICLE_ID' \
  --format markdown

python3 wechat-article-source/scripts/extract_wechat_article.py \
  --html /private/tmp/wechat_article.html \
  --format json

When direct URL fetching fails in the sandbox, fetch HTML with approved curl, save it to /private/tmp, then run the script with --html.

WeChat HTML Signals

Useful fields often appear in JavaScript variables:

• var msg_title
• var nickname
• var user_name
• var ct
• var msg_source_url
• var msg_cdn_url
• biz, mid, idx, sn

The article body is usually under id="js_content". Do not rely only on rendered screenshots when the HTML contains recoverable text.

Source-Checking Heuristics

Use focused searches, not broad query grids:

• exact article title
• quoted distinctive phrases from the article
• named product/project plus official-domain terms
• named documents, APIs, repositories, datasets, or companies
• public API endpoints mentioned in official docs, when safe and read-only

For each major claim, assign evidence strength:

• Strong: official document/API/page or original artifact confirms it
• Medium: credible third-party report confirms it, but primary source is unavailable
• Weak: only the WeChat article or screenshots/author claims support it
• Unverified: could not be checked from available public sources

Final Answer Shape

Keep the chat answer compact unless the user asks for a full report:

• article identity and publication metadata
• 5-8 key points
• source map with links
• verification notes and residual risk

When the article is promotional or experiential, say so plainly and separate product facts from the author's personal test results.