China Financial Claim Analyzer

Problem

Financial information often mixes facts, opinions, rumors, and policy narratives. Announcements, research notes, forum posts, and social media claims can all influence judgment. For investors, analysts, and regulators, the issue is not just whether information exists, but whether it is verifiable, whether the evidence is strong enough, and whether it may reflect market manipulation or selective disclosure.

How It Works

The skill processes a specific financial claim in stages:
- Claim decomposition: split complex statements into atomic items such as ENTITY, EVENT, STATE, FORECAST, CAUSATION, NARRATIVE, POLICY, and COMPETITIVE.
- Evidence grading: separate facts from opinions and assign A–E reliability levels to human and physical evidence, including regulatory documents, company announcements, media reports, and anonymous tips.
- Source validation: prioritize official sources such as CSRC, exchanges, CNINFO, and NBS, then cross-check with credible media and social platforms.
- Manipulation signals: monitor patterns like market-moving timing, hot-topic rotation, bulk distribution, and leak-first disclosure, and output a confidence rating with follow-up items.

Limits

It is best suited for validating specific claims, disclosures, research-note assertions, and rumors. It is not intended for macro research, quant signal generation, or full fundamental modeling, and final investment decisions still require human judgment.