Mbti Agent

What This Skill Does

The MBTI Agent skill is a specialized diagnostic tool designed specifically for non-interactive AI agents. Unlike standard human personality tests that rely on nuanced emotional responses, this skill provides a structured framework to evaluate the underlying tendencies of an AI model. It utilizes a static 70-question database to eliminate context drift and provides a batch-processing capability to ensure that the agent's personality analysis is consistent, reproducible, and mathematically grounded. By mapping machine-specific responses to the classic 16 MBTI personality types, it helps developers understand the core operational patterns of their AI agents, such as their preference for logical data processing (T) versus empathetic context (F), or their structural planning style (J) versus flexible adaptation (P).

Installation

To integrate this personality evaluation tool into your local OpenClaw environment, use the following terminal command:

clawhub install openclaw/skills/skills/guohongbin-git/mbti-agent

Once installed, ensure that your environment has Python 3 configured to allow the execution of the included scoring scripts. The system will create a workspace directory at ~/.openclaw/workspace/skills/mbti-agent/ which contains the core evaluation logic and database files.

Use Cases

• Base Personality Benchmarking: Determine the "native" personality of an LLM before applying custom system prompts.
• Agent Behavior Tuning: Verify if a specialized system prompt successfully shifts an agent's decision-making style toward a desired target personality type.
• Stability Analysis: Perform repeated tests across different conversation contexts to measure the variance in agent logic, helping developers assess if the agent remains reliable under load.

Example Prompts

1. "MBTI 测评：请帮我分析当前智能体的底层逻辑倾向。"
2. "我想知道我的人格类型，请启动人格测评工具并输出分析报告。"
3. "针对此任务上下文，请进行一次完整的人格测评，并将结果存储在 results 文件夹中。"

Tips & Limitations

To achieve the highest degree of accuracy, it is recommended to set your agent's Temperature parameter to 0 or a very low value during the evaluation. This minimizes randomness and ensures that the agent's choices are deterministic based on its core training. Note that the MBTI framework for AI is an interpretative tool rather than a psychological diagnosis; it serves as a proxy for understanding decision patterns. For the best results, run the test three times and compare the averages to ensure the agent's personality traits are stable and consistent.