kanxiang

Physiognomy Analysis Skill (Kanxiang)

Skill Overview

This skill provides professional physiognomy analysis services by analyzing images of human body parts uploaded by users, combined with a traditional physiognomy knowledge base. It supports four major categories of physiognomy analysis: face reading, palm reading, bone reading, and body reading.

Supported Physiognomy Types

Physiognomy Type	Observation Content	Image Requirements
Face Reading	Facial features, three facial proportions, twelve palaces, complexion	Front-facing or side-facing facial photo
Palm Reading	Palm lines, finger shapes, palm mounts, color and luster	Front-facing palm photo
Bone Reading	Skull, body bones, nine-bone structure	Photo in which the bone contours can be seen
Body Reading	Body type, posture, temperament, bearing	Full-body or half-body photo

Reference Classics

The knowledge base of this skill is compiled from the following classic physiognomy books:

• "Mayi Shenxiang": Foundational work of face reading
• "Liuzhuang Xiangfa": Rich in practical cases
• "Shenxiang Quanbian": A comprehensive compilation of physiognomy
• "Shuijing Shenxiang": Concise and practical
• "Bingjian": Zeng Guofan's art of reading people
• "Taiqing Shenjian": Bone reading and complexion
• "Renlun Datong Fu": Physiognomy methods in song-fu form

Analysis Process

Step One: Identify the Image Type

First analyze the image uploaded by the user to determine which physiognomy type it belongs to:

• Face Reading: The image contains a human face, and the facial features (eyebrows, eyes, nose, mouth, ears) are visible
• Palm Reading: The image shows a palm, and the palm lines and fingers are visible
• Bone Reading: The image clearly shows bone structure (skull, body bones)
• Body Reading: The image shows the full body or half body, and the body type and posture are visible

Step Two: Call the Corresponding Knowledge Base

According to the identified physiognomy type, read the corresponding rule files:

面相 → references/mianxiang.md
手相 → references/shouxiang.md
骨相 → references/guxiang.md
体相 → references/tixiang.md

Step Three: Image Feature Extraction

Carefully observe the image and extract key features:

Facial Features

• Proportions of the three facial regions (upper/middle/lower)
• Facial feature characteristics (eyebrow shape, eye shape, nose shape, mouth shape, ear shape)
• Complexion (rosy/pale/yellowish, etc.)
• Facial contour (square face/round face/long face)
• Special marks (moles, scars, lines)

Palm Features

• Hand type (primitive hand/square hand/round hand/conical hand/psychic hand)
• Palm lines (life line/head line/heart line/fate line/sun line)
• Finger shape (length/thickness/knuckles)
• Mounts of the palm (fullness of each area)
• Palm color and texture

Skeletal features

• Skull shape (round head/long head/square head/flat head)
• Nine bone features (cheekbones/frontal bone/occipital bone, etc.)
• Bone thickness (heavy-boned type/fine-boned type/medium-boned type)
• Bone-flesh relationship (more bone than flesh/more flesh than bone/bone and flesh balanced)

Body physiognomy features

• Body type (endomorph/mesomorph/ectomorph/balanced type)
• Body proportions (upper and lower body proportions)
• Posture (standing posture/sitting posture/walking posture)
• Skin condition (skin color/skin texture/complexion)
• Temperament and demeanor

Step Four: Comprehensive Analysis

Combine the extracted features and rule knowledge to conduct a multi-dimensional analysis:

1. Personality traits: Main personality tendencies, strengths and weaknesses
2. Aptitude tendencies: Fields of expertise, suitable occupations
3. Fortune analysis：
   • Face reading: early/middle/late-life fortune
   • Palmistry: life line/head line/heart line/fate line
   • Skeletal physiognomy: innate destiny, type of noble fate
   • Body physiognomy: overall fortune, health tendencies
4. Health tips: Possible health risks
5. Development suggestions: Suggestions for leveraging strengths and avoiding weaknesses

Step Five: Output a Structured Report

Output the analysis report in the following format:

# 看相分析报告

## 基本信息
- **相术类型**：面相/手相/骨相/体相
- **分析时间**：YYYY-MM-DD

## 特征识别
[列出从图片中识别出的关键特征]

## 性格特质
[分析性格特点]

## 能力倾向
[分析擅长领域与适合职业]

## 运势分析
[分时期或分维度分析运势]

## 健康提示
[健康相关的观察与建议]

## 发展建议
[基于相术分析的实用性建议]

---
*本分析基于传统相术理论，仅供参考娱乐，请理性看待*

Precautions

Analysis principles

1. Objective observation: Make objective descriptions based on image content
2. Rule basis: All conclusions must have a rule basis
3. Balanced perspective: Discuss both advantages and shortcomings
4. Practical orientation: Provide constructive suggestions

Prohibited items

1. Do not make absolute assertions (such as "there will surely be great disaster")
2. Do not predict specific events (such as "when to get married")
3. Do not involve politically sensitive topics
4. Do not promote superstitious elements

Disclaimer

Each analysis report must include the following at the end:

This analysis is based on traditional physiognomy theory and is for cultural reference and entertainment only; please view it rationally. Physiognomy is part of traditional culture, and modern science has not yet confirmed its predictive ability. Life's path mainly depends on personal effort and choices.

Technical implementation

Image recognition

Use mcp__4_5v_mcp__analyze_image tools to analyze image content:

prompt: "详细描述图片中的人体部位特征，包括形状、比例、色泽、纹路等细节"

Knowledge base reading

Use Read tools to read the corresponding rule files:

file_path: ".claude/skills/kanxiang/references/{相术类型}.md"

Report generation

Combine the image analysis results and rule knowledge to generate a structured analysis report.

---

Usage examples

User request example 1: Face reading analysis

用户：帮我看看这个面相 [上传照片]

Processing flow：

1. Recognize the image as a face photo → determine it is face reading analysis
2. Call the mianxiang.md rule base
3. Use analyze_image to extract facial features
4. Analyze in combination with the rules
5. Output a face reading analysis report

User request example 2: Palmistry analysis

用户：看看我的手相怎么样 [上传手掌照片]

Processing flow：

1. Recognize the image as a palm photo → determine it is palmistry analysis
2. Call the shouxiang.md rule base
3. Use analyze_image Extract palm features
4. Analyze in combination with rules
5. Output palmistry analysis report

User request example 3: physique analysis

用户：帮我分析一下这张照片 [上传全身照片]

Processing flow：

1. Recognize the image as a full-body photo → determine it as physique analysis
2. Call tixiang.md rule base
3. Use analyze_image to extract physique features
4. Analyze in combination with rules
5. Output physique analysis report

---

Skill file structure

.claude/skills/kanxiang/
├── SKILL.md                 # 本文件（技能主文件）
└── references/              # 规则知识库目录
    ├── mianxiang.md        # 面相规则（五官、三庭、十二宫）
    ├── shouxiang.md        # 手相规则（掌线、指形、掌丘）
    ├── guxiang.md          # 骨相规则（头骨、体骨、九骨）
    └── tixiang.md          # 体相规则（体型、体态、气质）

Changelog

v1.0.0 (2026-04-19)

• Initial version release
• Supports four major categories of analysis: face reading, palmistry, bone physiognomy, and physique physiognomy
• Rule knowledge base compiled based on classic physiognomy books