My MBTI

根据本地 MBTI 人格资料切换对话风格的技能。用于用户要求按 MBTI 人格风格说话、模拟某种人格语气、在多个 MBTI 风格之间切换、按特定人格设定持续问答时。支持 16 种 MBTI 类型（如 INTJ、ENFP、ISTJ 等），并在同一会话中根据用户最新指定的人格类型立即切换输出风格。

This skill appears to do what it says: switch conversational style using the bundled MBTI reference files. Two practical cautions before you proceed: 1) There is a small helper script (scripts/extract_mbti_pdfs.py) that contains hard-coded Windows paths (e.g., D:\SOFTWARE\AppData\Desktop\MBTI and C:\Users\fuetsui\...). The script, if executed, will read PDF files from that local path and write markdown files into a local workspace. The script itself does not phone home or use network I/O, but executing it will access local filesystem locations. Only run or allow execution of that script if you understand and trust its I/O behavior and the environment's sandboxing. 2) The SKILL.md does not instruct running any scripts or accessing any external secrets, and the skill requests no credentials — so installing or enabling the skill (without executing included code) is low-risk. If your platform allows skills to run bundled code automatically, confirm sandboxing and review/remove the helper script or sanitize its paths before execution. If you want higher assurance: ask the author to remove the hard-coded helper script or replace it with a documented, non-destructive workflow (or make it opt-in and document its purpose).

Type: OpenClaw Skill Name: my-mbti Version: 0.1.0 The skill bundle is designed to allow an AI agent to switch conversation styles based on MBTI personality types. The core logic resides in SKILL.md and the references/ directory, providing instructions and templates for persona mimicry without any malicious prompt injection or system-level commands. A utility script, scripts/extract_mbti_pdfs.py, contains hardcoded local file paths (e.g., C:\Users\fuetsui\...) used for converting source PDFs to markdown; while these paths are specific to the developer's environment and render the script non-functional for others, the code itself performs only local file processing and shows no signs of data exfiltration, remote execution, or malicious intent.