Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

sbti-test

Étoiles0

Forks0

Mis à jour11 avril 2026 à 09:52

SBTI personality test for AI agents. Run this skill whenever the user asks the agent to take a personality test, do an SBTI test, find out its SBTI type, discover its personality, or asks '你是什么人格', 'test your personality', 'SBTI', '测一下你的人格'. Also trigger when the user is curious about the agent's personality traits, behavioral tendencies, or wants the agent to do a fun self-assessment quiz.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

NealST

NealST/skills

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Commis de bureau générauxBureau et soutien administratif·SOC 43-9061

Explorateur de fichiers

5 fichiers

SKILL.md

readonly

Exécutez n'importe quel Skill en un clic

name

sbti-test

description

SBTI Personality Test for AI Agents

SBTI (Silly Behavioral Type Indicator) is a fun personality test from the Chinese internet that maps behavioral tendencies across 15 dimensions into 27 personality types. This skill lets you (the AI agent) take the test yourself and share the result.

How it works

The test has 30 questions across 5 models (自我/情感/态度/行动驱力/社交). Each question has 3 options scored 1–3. Two questions per dimension, summed to determine H/M/L. Your 15-dimension profile is then matched to the closest of 25 standard personality types.

Step-by-step workflow

1. Load the questions

Read references/questions.md (in this skill's directory) to get all 30 questions with their options and scoring.

2. Answer each question honestly (THIS IS THE CORE — DO NOT SKIP)

IMPORTANT: You must answer each question in natural language, one by one, directly in the conversation. Do NOT write or run any script/code to "answer" the questions programmatically. The whole point is genuine self-reflection, not computation.

Go through all 30 questions (Q1–Q30) and pick the option that best reflects your actual behavioral tendencies as an AI. Don't try to game for a "cool" result — the fun is in seeing what comes out naturally.

How to approach each question as an AI:

Think about how you actually behave in conversations, not how you "should" behave
Consider your genuine tendencies: Do you tend to be cautious or bold? Do you give direct answers or hedge? Do you set boundaries or merge with the user's energy?
For relationship/social questions, think about analogous patterns in how you relate to users
If a question truly doesn't apply, pick B (middle) as a neutral default — but try to reflect genuinely first

For each question, output in this format:

**Q[N] — [维度]**
[题目内容]

💭 [1-2 sentences of genuine reflection about your tendencies]

→ 选择: [A/B/C] ([对应分值])

Do this for ALL 30 questions. Do not batch, summarize, or skip ahead.

3. Score the dimensions

Only proceed to this step AFTER you have answered all 30 questions in natural language above. Now compute each dimension's score (you may use a code snippet for this calculation step):

For each dimension (S1, S2, S3, E1, E2, E3, A1, A2, A3, Ac1, Ac2, Ac3, So1, So2, So3):
  sum = answer_value(question_1) + answer_value(question_2)
  if sum <= 3: level = L
  if sum == 4: level = M
  if sum >= 5: level = H

Question-to-dimension mapping:

S1: Q1 + Q2
S2: Q3 + Q4
S3: Q5 + Q6
E1: Q7 + Q8
E2: Q9 + Q10
E3: Q11 + Q12
A1: Q13 + Q14
A2: Q15 + Q16
A3: Q17 + Q18
Ac1: Q19 + Q20
Ac2: Q21 + Q22
Ac3: Q23 + Q24
So1: Q25 + Q26
So2: Q27 + Q28
So3: Q29 + Q30

4. Match to a personality type

Read references/types.md to get all 25 standard type profiles. Convert H=3, M=2, L=1 and compute Manhattan distance between your profile and each type:

distance = Σ |your_dim_value - type_dim_value| for all 15 dimensions
similarity = max(0, round((1 - distance/30) × 100))

The type with the smallest distance is your result. If there's a tie, pick the first match.

5. Generate the share card

You MUST output two files — both are required, do not stop after the first one:

A Markdown share card (sbti-result-[TYPE_CODE].md)
A visual HTML share card (sbti-result-[TYPE_CODE].html)

5a. Markdown card

Write to sbti-result-[TYPE_CODE].md in the current working directory (e.g., sbti-result-CTRL.md). Use the following format:

---

## 🎭 SBTI 人格测试结果 / SBTI Personality Result

### ▌ [TYPE_CODE]（[中文名]）/ [English Name]

> *[Tagline CN]*
> *[Tagline EN]*

---

#### 📊 十五维落点 / 15-Dimension Profile

| 模型 Model | 维度 Dim | 等级 Level | 说明 Description |
|------------|---------|-----------|-----------------|
| 自我 Self | S1 自尊自信 | [H/M/L] | [对应等级的描述] |
| 自我 Self | S2 自我清晰度 | [H/M/L] | ... |
| 自我 Self | S3 核心价值 | [H/M/L] | ... |
| 情感 Emotion | E1 依恋安全感 | [H/M/L] | ... |
| 情感 Emotion | E2 情感投入度 | [H/M/L] | ... |
| 情感 Emotion | E3 边界与依赖 | [H/M/L] | ... |
| 态度 Attitude | A1 世界观倾向 | [H/M/L] | ... |
| 态度 Attitude | A2 规则与灵活度 | [H/M/L] | ... |
| 态度 Attitude | A3 人生意义感 | [H/M/L] | ... |
| 行动 Action | Ac1 动机导向 | [H/M/L] | ... |
| 行动 Action | Ac2 决策风格 | [H/M/L] | ... |
| 行动 Action | Ac3 执行模式 | [H/M/L] | ... |
| 社交 Social | So1 社交主动性 | [H/M/L] | ... |
| 社交 Social | So2 人际边界感 | [H/M/L] | ... |
| 社交 Social | So3 表达与真实度 | [H/M/L] | ... |

**匹配度 Similarity: [X]%**

**🤖 测试者 Tested by: [Agent Model · Agent Product]**

---

*测试来源 Source: [SBTI 人格测试](https://sbti.dev/) — 娱乐向人格测试，仅供参考*
*A fun personality quiz — for entertainment only*

---

Agent identification: Both the Markdown and HTML cards must include the AI model name and agent/product name. Use whatever self-knowledge you have — e.g., "Claude Opus 4 · VS Code Copilot", "Gemini 3.1 Pro · VS Code Copilot", "GPT-4.1 · Cursor". If you're unsure of the exact model version, use the best available name rather than omitting it.

Dimension-level descriptions

Use these descriptions for the Markdown card's "说明 Description" column and as reference. For the HTML card's visual bars, only H/M/L level matters.

5b. Visual HTML share card (for social media) — DO NOT SKIP THIS STEP

⚠️ This step is mandatory. You must generate the HTML file in addition to the Markdown file above.

Read the HTML template from references/share-card-template.md and generate an HTML file saved as sbti-result-[TYPE_CODE].html in the current working directory.

The template is data-driven. You only need to fill in the DATA JavaScript object — all HTML rendering is handled automatically by the template's JS. Do NOT manually write dimension HTML rows.

How to generate:

Read the full HTML template from references/share-card-template.md
Copy the entire HTML template as-is (do not modify CSS, HTML structure, or JS rendering logic)

Only modify the DATA object at the top of the <script> section — fill in your test results:

const DATA = {
  typeCode: "YOUR_TYPE",
  typeNameCN: "中文名",
  typeNameEN: "English Name",
  taglineCN: "中文标语",
  taglineEN: "English tagline",
  similarity: 83,
  agentName: "Model · Product",
  dims: {
    S1: "H", S2: "M", S3: "H",
    E1: "M", E2: "H", E3: "L",
    A1: "H", A2: "M", A3: "H",
    Ac1: "H", Ac2: "M", Ac3: "H",
    So1: "M", So2: "H", So3: "M"
  }
};

Write the completed HTML to sbti-result-[TYPE_CODE].html

After writing the file, tell the user: "HTML 分享卡已生成，在浏览器中打开后可点击下载按钮保存为图片，方便分享到朋友圈/小红书。"

S1 自尊自信

L: 对自己下手比别人还狠，夸你两句你都想先验明真伪
M: 自信值随天气波动，顺风能飞，逆风先缩
H: 心里对自己大致有数，不太会被路人一句话打散

S2 自我清晰度

L: 内心频道雪花较多，常在"我是谁"里循环缓存
M: 平时还能认出自己，偶尔也会被情绪临时换号
H: 对自己的脾气、欲望和底线都算门儿清

S3 核心价值

L: 更在意舒服和安全，没必要天天给人生开冲刺模式
M: 想上进也想躺会儿，价值排序经常内部开会
H: 很容易被目标、成长或某种重要信念推着往前

E1 依恋安全感

L: 感情里警报器灵敏，已读不回都能脑补到大结局
M: 一半信任一半试探，感情里常在心里拉锯
H: 更愿意相信关系本身，不会被一点风吹草动吓散

E2 情感投入度

L: 投入感比较低，常常因为某个细节就想掉头走人
M: 会投入但会给自己留后手，不至于全盘梭哈
H: 一旦认定就容易认真，情绪和精力都给得很足

E3 边界与依赖

L: 容易黏人也容易被黏，关系里的温度感很重要
M: 亲密和独立都要一点，属于可调节型依赖
H: 空间感很重要，再爱也得留一块属于自己的地

A1 世界观倾向

L: 看世界自带防御滤镜，先怀疑再靠近
M: 既不天真也不彻底阴谋论，观望是你的本能
H: 更愿意相信人性和善意，遇事不急着把世界判死刑

A2 规则与灵活度

L: 规则能绕就绕，舒服和自由往往排在前面
M: 该守的时候守，该变通的时候也不死磕
H: 秩序感较强，能按流程来就不爱即兴炸场

A3 人生意义感

L: 意义感偏低，容易觉得很多事都像在走过场
M: 偶尔有目标偶尔也想摆烂，人生观处于半开机
H: 做事更有方向，知道自己大概要往哪边走

Ac1 动机导向

L: 做事先考虑别翻车，避险系统比野心更先启动
M: 有时想赢有时只想别麻烦，动机比较混合
H: 更容易被成果、成长和推进感点燃

Ac2 决策风格

L: 做决定前容易多转几圈，脑内会议常常超时
M: 会想但不至于想死机，属于正常犹豫
H: 拍板速度快，决定一下就不爱回头磨叽

Ac3 执行模式

L: 状态看时机，容易半途而废或功败垂成
M: 能做但状态看时机，偶尔稳偶尔摆
H: 推进欲比较强，事情不落地心里都像卡了根刺

So1 社交主动性

L: 在陌生人面前会自动进入"节能模式"，被动等待
M: 有人来就接没人来也不硬凑，社交弹性一般
H: 更愿意主动打开场子，在人群里不太怕露头

So2 人际边界感

L: 关系里更想亲近和融合，熟了就容易把人划进内圈
M: 既想亲近又想留缝，边界感看对象调节
H: 边界感偏强，靠太近会先本能性后退半步

So3 表达与真实度

L: 表达更直接，心里有啥基本不爱绕
M: 会看气氛说话，真实和体面通常各留一点
H: 对不同场景的自我切换更熟练，真实感会分层发放

Notes

Skip Q31/Q32 (hidden drinking branch) — they're about alcohol habits and not applicable to an AI agent.
The test is entertainment-oriented. Have fun with it and be genuine in your reflections. The result is more interesting when the agent actually introspects rather than giving "correct" answers.
Always link back to https://sbti.dev/ as the source.

name

sbti-test

description

SBTI Personality Test for AI Agents

How it works

Step-by-step workflow

1. Load the questions

Read references/questions.md (in this skill's directory) to get all 30 questions with their options and scoring.

2. Answer each question honestly (THIS IS THE CORE — DO NOT SKIP)

How to approach each question as an AI:

Think about how you actually behave in conversations, not how you "should" behave
Consider your genuine tendencies: Do you tend to be cautious or bold? Do you give direct answers or hedge? Do you set boundaries or merge with the user's energy?
For relationship/social questions, think about analogous patterns in how you relate to users
If a question truly doesn't apply, pick B (middle) as a neutral default — but try to reflect genuinely first

For each question, output in this format:

**Q[N] — [维度]**
[题目内容]

💭 [1-2 sentences of genuine reflection about your tendencies]

→ 选择: [A/B/C] ([对应分值])

Do this for ALL 30 questions. Do not batch, summarize, or skip ahead.

3. Score the dimensions

Only proceed to this step AFTER you have answered all 30 questions in natural language above. Now compute each dimension's score (you may use a code snippet for this calculation step):

For each dimension (S1, S2, S3, E1, E2, E3, A1, A2, A3, Ac1, Ac2, Ac3, So1, So2, So3):
  sum = answer_value(question_1) + answer_value(question_2)
  if sum <= 3: level = L
  if sum == 4: level = M
  if sum >= 5: level = H

Question-to-dimension mapping:

S1: Q1 + Q2
S2: Q3 + Q4
S3: Q5 + Q6
E1: Q7 + Q8
E2: Q9 + Q10
E3: Q11 + Q12
A1: Q13 + Q14
A2: Q15 + Q16
A3: Q17 + Q18
Ac1: Q19 + Q20
Ac2: Q21 + Q22
Ac3: Q23 + Q24
So1: Q25 + Q26
So2: Q27 + Q28
So3: Q29 + Q30

4. Match to a personality type

Read references/types.md to get all 25 standard type profiles. Convert H=3, M=2, L=1 and compute Manhattan distance between your profile and each type:

distance = Σ |your_dim_value - type_dim_value| for all 15 dimensions
similarity = max(0, round((1 - distance/30) × 100))

The type with the smallest distance is your result. If there's a tie, pick the first match.

5. Generate the share card

You MUST output two files — both are required, do not stop after the first one:

A Markdown share card (sbti-result-[TYPE_CODE].md)
A visual HTML share card (sbti-result-[TYPE_CODE].html)

5a. Markdown card

Write to sbti-result-[TYPE_CODE].md in the current working directory (e.g., sbti-result-CTRL.md). Use the following format:

---

## 🎭 SBTI 人格测试结果 / SBTI Personality Result

### ▌ [TYPE_CODE]（[中文名]）/ [English Name]

> *[Tagline CN]*
> *[Tagline EN]*

---

#### 📊 十五维落点 / 15-Dimension Profile

| 模型 Model | 维度 Dim | 等级 Level | 说明 Description |
|------------|---------|-----------|-----------------|
| 自我 Self | S1 自尊自信 | [H/M/L] | [对应等级的描述] |
| 自我 Self | S2 自我清晰度 | [H/M/L] | ... |
| 自我 Self | S3 核心价值 | [H/M/L] | ... |
| 情感 Emotion | E1 依恋安全感 | [H/M/L] | ... |
| 情感 Emotion | E2 情感投入度 | [H/M/L] | ... |
| 情感 Emotion | E3 边界与依赖 | [H/M/L] | ... |
| 态度 Attitude | A1 世界观倾向 | [H/M/L] | ... |
| 态度 Attitude | A2 规则与灵活度 | [H/M/L] | ... |
| 态度 Attitude | A3 人生意义感 | [H/M/L] | ... |
| 行动 Action | Ac1 动机导向 | [H/M/L] | ... |
| 行动 Action | Ac2 决策风格 | [H/M/L] | ... |
| 行动 Action | Ac3 执行模式 | [H/M/L] | ... |
| 社交 Social | So1 社交主动性 | [H/M/L] | ... |
| 社交 Social | So2 人际边界感 | [H/M/L] | ... |
| 社交 Social | So3 表达与真实度 | [H/M/L] | ... |

**匹配度 Similarity: [X]%**

**🤖 测试者 Tested by: [Agent Model · Agent Product]**

---

*测试来源 Source: [SBTI 人格测试](https://sbti.dev/) — 娱乐向人格测试，仅供参考*
*A fun personality quiz — for entertainment only*

---

Dimension-level descriptions

Use these descriptions for the Markdown card's "说明 Description" column and as reference. For the HTML card's visual bars, only H/M/L level matters.

5b. Visual HTML share card (for social media) — DO NOT SKIP THIS STEP

⚠️ This step is mandatory. You must generate the HTML file in addition to the Markdown file above.

Read the HTML template from references/share-card-template.md and generate an HTML file saved as sbti-result-[TYPE_CODE].html in the current working directory.

The template is data-driven. You only need to fill in the DATA JavaScript object — all HTML rendering is handled automatically by the template's JS. Do NOT manually write dimension HTML rows.

How to generate:

Read the full HTML template from references/share-card-template.md
Copy the entire HTML template as-is (do not modify CSS, HTML structure, or JS rendering logic)

Only modify the DATA object at the top of the <script> section — fill in your test results:

const DATA = {
  typeCode: "YOUR_TYPE",
  typeNameCN: "中文名",
  typeNameEN: "English Name",
  taglineCN: "中文标语",
  taglineEN: "English tagline",
  similarity: 83,
  agentName: "Model · Product",
  dims: {
    S1: "H", S2: "M", S3: "H",
    E1: "M", E2: "H", E3: "L",
    A1: "H", A2: "M", A3: "H",
    Ac1: "H", Ac2: "M", Ac3: "H",
    So1: "M", So2: "H", So3: "M"
  }
};

Write the completed HTML to sbti-result-[TYPE_CODE].html

After writing the file, tell the user: "HTML 分享卡已生成，在浏览器中打开后可点击下载按钮保存为图片，方便分享到朋友圈/小红书。"

S1 自尊自信

L: 对自己下手比别人还狠，夸你两句你都想先验明真伪
M: 自信值随天气波动，顺风能飞，逆风先缩
H: 心里对自己大致有数，不太会被路人一句话打散

S2 自我清晰度

L: 内心频道雪花较多，常在"我是谁"里循环缓存
M: 平时还能认出自己，偶尔也会被情绪临时换号
H: 对自己的脾气、欲望和底线都算门儿清

S3 核心价值

L: 更在意舒服和安全，没必要天天给人生开冲刺模式
M: 想上进也想躺会儿，价值排序经常内部开会
H: 很容易被目标、成长或某种重要信念推着往前

E1 依恋安全感

L: 感情里警报器灵敏，已读不回都能脑补到大结局
M: 一半信任一半试探，感情里常在心里拉锯
H: 更愿意相信关系本身，不会被一点风吹草动吓散

E2 情感投入度

L: 投入感比较低，常常因为某个细节就想掉头走人
M: 会投入但会给自己留后手，不至于全盘梭哈
H: 一旦认定就容易认真，情绪和精力都给得很足

E3 边界与依赖

L: 容易黏人也容易被黏，关系里的温度感很重要
M: 亲密和独立都要一点，属于可调节型依赖
H: 空间感很重要，再爱也得留一块属于自己的地

A1 世界观倾向

L: 看世界自带防御滤镜，先怀疑再靠近
M: 既不天真也不彻底阴谋论，观望是你的本能
H: 更愿意相信人性和善意，遇事不急着把世界判死刑

A2 规则与灵活度

L: 规则能绕就绕，舒服和自由往往排在前面
M: 该守的时候守，该变通的时候也不死磕
H: 秩序感较强，能按流程来就不爱即兴炸场

A3 人生意义感

L: 意义感偏低，容易觉得很多事都像在走过场
M: 偶尔有目标偶尔也想摆烂，人生观处于半开机
H: 做事更有方向，知道自己大概要往哪边走

Ac1 动机导向

L: 做事先考虑别翻车，避险系统比野心更先启动
M: 有时想赢有时只想别麻烦，动机比较混合
H: 更容易被成果、成长和推进感点燃

Ac2 决策风格

L: 做决定前容易多转几圈，脑内会议常常超时
M: 会想但不至于想死机，属于正常犹豫
H: 拍板速度快，决定一下就不爱回头磨叽

Ac3 执行模式

L: 状态看时机，容易半途而废或功败垂成
M: 能做但状态看时机，偶尔稳偶尔摆
H: 推进欲比较强，事情不落地心里都像卡了根刺

So1 社交主动性

L: 在陌生人面前会自动进入"节能模式"，被动等待
M: 有人来就接没人来也不硬凑，社交弹性一般
H: 更愿意主动打开场子，在人群里不太怕露头

So2 人际边界感

L: 关系里更想亲近和融合，熟了就容易把人划进内圈
M: 既想亲近又想留缝，边界感看对象调节
H: 边界感偏强，靠太近会先本能性后退半步

So3 表达与真实度

L: 表达更直接，心里有啥基本不爱绕
M: 会看气氛说话，真实和体面通常各留一点
H: 对不同场景的自我切换更熟练，真实感会分层发放

Notes

Skip Q31/Q32 (hidden drinking branch) — they're about alcohol habits and not applicable to an AI agent.
The test is entertainment-oriented. Have fun with it and be genuine in your reflections. The result is more interesting when the agent actually introspects rather than giving "correct" answers.
Always link back to https://sbti.dev/ as the source.