WeChat Stock KOL Distillation Analysis Expert

Problem

WeChat stock KOL posts are often fragmented, tonal, and hard to reuse as a repeatable analysis process. A plain LLM rewrite may still treat market data as fact or imitate surface phrasing without replicating the underlying framework.

How It Works

The skill distills multiple KOL articles into structured profiles covering trading system, market judgment, expression style, content depth, interaction patterns, and topic maps. For each request, it loads the relevant profile, then uses WebSearch to collect real-time market data such as limit-up/limit-down counts, turnover, northbound funds, and main capital flow, with dual-source validation where possible.

• Multi-KOL comparison: generates different style-based views on the same ticker or sector and highlights consensus and divergence.
• Review and selection: supports daily review, sector funnel, short-term selection, portfolio review, and earnings commentary.
• Audit gate: surface features and reasoning-layer features each carry 50% weight, and output requires a score above 70 with explicit data-source annotations.

Boundaries

Market data must be collected in real time and must not be fabricated by the LLM. The generated content is an AI style simulation and is not investment advice. New-article sync triggers supplemental distillation to update profiles and flag style drift.