How the Soul Inquiry Method Works in Zhangxuefeng-skill

The Soul Inquiry Method (灵魂追问法) is a mandatory three-question checkpoint that forces the zhangxuefeng-skill to collect decisive personal variables—exam score, province, and family occupation—before generating any recommendation, ensuring every answer is grounded in the user's specific economic and geographic context rather than generic advice.

The Soul Inquiry Method is the foundational decision heuristic in the alchaincyf/zhangxuefeng-skill repository. As one of eight decision heuristics listed in the project's documentation, it transforms the skill from a static knowledge base into an interactive advisor that mirrors Zhang Xuefeng's rapid-fire live-stream questioning style. According to the source files README.md and SKILL.md, this method ensures that no recommendation is made until the engine has gathered the three most critical variables that distinguish "有矿" (wealthy) families from "没矿" (ordinary) ones.

What Is the Soul Inquiry Method?

The Soul Inquiry Method (灵魂追问法) ranks among the "8条决策启发式" (eight decision heuristics) defined in [README.md](https://github.com/alchaincyf/zhangxuefeng-skill/blob/main/README.md) at line 40. It functions as the skill's primary data-gathering protocol, designed to quickly lock in a solution by extracting high-impact personal data before any analytical model is applied.

Unlike generic advisory systems that immediately fetch employment statistics, the Soul Inquiry Method enforces a strict sequential workflow. The skill must first execute the "🔴 CHECKPOINT · 开口前三问" (First Three Questions Checkpoint) documented in [SKILL.md](https://github.com/alchaincyf/zhangxuefeng-skill/blob/main/SKILL.md) lines 80-86. If these three critical questions remain unanswered, the engine automatically falls back to the Soul Inquiry routine before proceeding to data retrieval or recommendation generation.

The Three Critical Variables of the Soul Inquiry

The method revolves around three specific queries that Zhang Xuefeng uses to separate decision-making contexts. These questions are hardcoded into the skill's operational workflow:

• Score: "你孩子多少分？" (What is your child's exam score?)
• Location: "哪个省的？" (Which province are you in?)
• Family Background: "家里做什么的？" (What does your family do for a living?)

These three variables capture the essential economic and geographic context required to apply Zhang Xuefeng's analytical models, including the 社会筛子论 (Social Sieve Theory) and 就业倒推法 (Employment Backcasting Method).

Implementation in the Skill Pipeline

The Soul Inquiry Method is embedded directly into the skill's answer pipeline as Step 1 of a three-step process. According to SKILL.md, the workflow follows this strict sequence:

1. Step 1: Soul Inquiry — Verify the three questions have been answered
2. Step 2: Data Retrieval — Fetch latest employment statistics and median earnings via web search
3. Step 3: Model Application — Apply relevant mind-models and render the final verdict

The checkpoint logic ensures that the engine cannot bypass the Soul Inquiry. If is_personal_choice(user_input) returns true but the context object lacks score, province, or family_bg, the system triggers the inquiry routine before allowing the pipeline to advance to Step 2.

Code Example: The Decision Pipeline

The following Python pseudocode from the skill's internal logic demonstrates how the Soul Inquiry Method gates the recommendation engine:

def handle_question(user_input):
    if is_personal_choice(user_input):
        # Step 1 – Soul Inquiry Checkpoint

        ask("你孩子多少分？")
        ask("哪个省的？")
        ask("家里做什么的？")
        
        # Collect answers into context

        ctx = {
            'score': get_score(),
            'province': get_province(),
            'family_bg': get_family_background()
        }
        
        # Step 2 – Fetch employment data

        data = fetch_employment_stats(ctx['province'], ctx['score'])
        
        # Step 3 – Apply mind-models and render

        return render_answer(ctx, data)

In this implementation, the function explicitly halts execution to gather the three critical variables, ensuring that subsequent employment data queries are filtered by the user's actual geographic and economic constraints.

Real-World Usage Patterns

You can install the skill in Agent-Skills compatible runtimes using npm:

# Install the skill

npx skills add alchaincyf/zhangxuefeng-skill

Once installed, the method triggers automatically on personal choice queries. For example:

User: 我孩子高考 560 分，河南，家里没有金融背景，应该选什么专业？

The skill recognizes this as a personal choice question and immediately validates the three variables. If any are missing, it initiates the Soul Inquiry:

Agent: 先问几个关键信息——你现在的分数是多少？（如果有成绩）
Agent: 你所在的省份是？
Agent: 你家里主要做什么行业？

After collecting these answers, the engine pulls 2026 AI岗位 employment rates and median salary data, applies the 就业倒推法 model, and returns a decisive statement such as "这条路能走，但你家没矿的话先保证一年内能跑到 20 万以上才安全".

Why the Method Works

The Soul Inquiry Method mimics Zhang Xuefeng's live-stream technique of "快速追问" (rapid questioning) to "锁定最优方案" (lock in the optimal solution). By iteratively narrowing the decision space through these three high-impact variables, the method eliminates generic advice that ignores economic reality.

Real-world examples of this questioning pattern are documented in [references/research/02-conversations.md](https://github.com/alchaincyf/zhangxuefeng-skill/blob/main/references/research/02-conversations.md), which captures Zhang's actual live-stream interactions. The result is a concise, data-driven verdict that feels personal and instantly actionable because it is rooted in the user's concrete context rather than abstract trends.

Summary

• The Soul Inquiry Method is one of eight decision heuristics defined in README.md line 40, serving as the mandatory first step in the skill's recommendation pipeline.
• It enforces a three-question checkpoint (score, location, family background) documented in SKILL.md lines 80-86 before any data retrieval occurs.
• The method gates the workflow via the handle_question function, ensuring the engine collects ctx variables before applying employment models like 就业倒推法.
• By distinguishing between "有矿" and "没矿" family contexts, it transforms generic advice into specific, economically grounded recommendations.
• The implementation mirrors Zhang Xuefeng's documented live-stream questioning style, creating an interactive advisor rather than a static search tool.

Frequently Asked Questions

What are the exact three questions asked by the Soul Inquiry Method?

The method asks: "你孩子多少分？" (What is the exam score?), "哪个省的？" (Which province?), and "家里做什么的？" (What is the family's occupation?). These three questions are hardcoded into the checkpoint logic in SKILL.md lines 80-86 and are designed to capture the essential variables needed to filter employment data and apply Zhang Xuefeng's economic models.

Where is the Soul Inquiry Method implemented in the codebase?

The method is defined in the heuristic list in [README.md](https://github.com/alchaincyf/zhangxuefeng-skill/blob/main/README.md) line 40 and implemented in the workflow checkpoint section of [SKILL.md](https://github.com/alchaincyf/zhangxuefeng-skill/blob/main/SKILL.md) lines 80-86. The operational logic appears in the handle_question function, which checks is_personal_choice() before triggering the inquiry routine.

How does the Soul Inquiry Method differ from generic career advice systems?

Generic systems typically query employment databases immediately, providing one-size-fits-all answers. The Soul Inquiry Method forces a context-first approach by requiring exam scores, geographic location, and family economic background before accessing any employment statistics. This ensures recommendations account for regional job markets and family economic constraints, distinguishing between "有矿" (wealthy) and "没矿" (ordinary) scenarios.

Can the Soul Inquiry Method be customized or extended?

While the current implementation in SKILL.md hardcodes the three specific questions to match Zhang Xuefeng's methodology, the checkpoint structure (🔴 CHECKPOINT · 开口前三问) is modular. Developers could extend the ctx object in the handle_question function to include additional variables, though doing so would diverge from the validated "8条决策启发式" model documented in the repository.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}