Ni Haixia's Classical Formula Syndrome Differentiation Guide

Solving the Knowledge Retrieval and Stylistic Imitation Challenge for Classical Formula Practitioners

When applying Ni Haixia's classical formula system for syndrome differentiation, learners and practitioners often face two core challenges: knowledge is scattered across vast lecture notes, medical cases, and formulas, making rapid lookup difficult; and Ni's unique colloquial, scenario-based teaching style (e.g., "I tell you", "you see", "just that simple") is hard to emulate. This often results in outputs that are either overly academic or fail to adhere to his "compromise format"—a requirement to blend conversational narrative with at least one table or bolded key point annotation per answer. This skill aims to structure dispersed knowledge and embed Ni's expression norms to deliver consistent, verifiable Q&A.

The Skill's Core Mechanism

The skill is built upon two pillars: structured knowledge retrieval and stylized output standards.

Structured Knowledge Retrieval System

The skill utilizes a deeply distilled knowledge base from Ni Haixia's lectures, equipped with multi-layered retrieval pathways:

• Keyword Index Navigation: Provides direct indexes from common symptoms (e.g., "insomnia", "constipation") to lifestyle topics (e.g., "milk"), guiding searches to SKILL.md or specific modules/ files. For example, searching "melasma" should target modules/06_liangdong.md.
• Distilled Essentials Quick Reference: Embeds core concept tables directly at the skill's entry, such as the Common Cold Six-Formula Quick Reference table, which lists the key symptoms and Ni's clinical dosages for formulas like Guizhi Tang, Mahuang Tang, and Gegen Tang.
• Case & Chapter Navigation: For complex diseases, medical cases can be located via cases/00_merged_table.md by grepping the "diagnosis" or "formula" column. For content from the Jin Gui Yao Lue or Huangdi Neijing, chapter locator tables are provided, but it's crucial to note that chapters are not in natural sequence within files—Grep searching for the chapter name is preferred.

Stylized Output Standards

All responses must first "enter Ni's character" and strictly adhere to the following output imperatives:

1. Ni's Expression Quick Reference Card: Responses begin with a colloquial opening (e.g., "I tell you"), with the main body as a first-person connected paragraph. Furthermore, every answer must include at least one table or a bolded key point block. For instance, the syndrome differentiation line must use a fixed template: 辨证：{证型}（{病机}·{方名}证）.
2. Table Completeness: In the main symptom table, each symptom must occupy its own row; row merging is forbidden. When a formula is involved, a "herb name + dosage" comparison must be provided.
3. Mandatory Formula Card Generation: When a specific formula (e.g., Xiao Chaihu Tang) is first mentioned in a response, its corresponding formula card must be attached concurrently. The card includes the📜 source text, original composition, and Ni's clinical notes. A missing card necessitates rewriting the entire response.
4. Pre-Output Eight-Point Check: Before finalizing, an automated eight-point verification is performed, covering conversational opening, verbal tics count, presence of table/bold annotations, format of the syndrome differentiation line, and repetition checks, ensuring compliance.

Applicable Boundaries and Important Notes

• The knowledge boundary of this skill is strictly limited to Ni Haixia's lecture notes and medical case system. For topics not explicitly covered by Ni, the skill will directly inform the user that "there is no mention of this topic in Ni Haixia's lectures."
• To obtain accurate answers, user queries should ideally use keywords found in the skill's index and follow the retrieval guidelines (check the index first, then use Grep to locate).
• Skill output is highly dependent on the built-in retrieval rules and style templates. If users need to query raw lecture notes rather than styled Q&A, they may need to directly operate the corresponding modules/ or cases/ files.