china source intelligence

The problem

English-language research on Chinese markets is often shallow, relying on translated snippets, outdated reports, or unreliable social media posts. Surface-level searches miss primary regulatory filings, official circulars, and industry-specific data locked behind the Great Firewall or specific Chinese-language terminology.

What it does

• Generates bilingual query matrices using Simplified and Traditional Chinese to penetrate localized search engines like Baidu and Sogou.
• Automates extraction of public Chinese web pages while maintaining metadata, original encoding, and source-tier classification.
• Builds a JSONL evidence ledger that pairs original Chinese text with faithful English translations for every load-bearing claim.
• Performs multi-layered verification across regulators, ministries, courts, and exchanges to ensure accuracy of legal and financial data.
• Identifies and reports material contradictions rather than averaging conflicting data points.

Frameworks & tools

Python scripts for query building and page extraction, JSONL for evidence tracking, and source-specific discovery protocols for WeChat, Zhihu, and Bilibili.

Why this beats prompting it yourself

Standard LLM prompts often hallucinate Chinese regulatory status or fail to account for modal nuances in Chinese text like "proposed" vs "ordered." This skill enforces a rigorous audit trail, requiring Tier 1 evidence for legal claims and tracing every screenshot or article back to a primary document number.

Use cases

• Due diligence on mainland Chinese corporate entities and their legal histories.
• Deep-dive investigations into Chinese government policy, standards, and regulatory shifts.
• Product and supply chain research using local procurement records and technical communities.
• Public sentiment analysis via WeChat and Zhihu to gauge local reaction to global brands.

Known limitations

Cannot bypass logins, paywalls, or robots.txt restrictions. Performance on dynamic JavaScript-heavy pages may require the documented fallbacks in the source discovery protocol.