SBTI Personality Test

Problem Solved

The SBTI test at sbti.unun.dev is built as a web form, so users have to click through questions, wait for page updates, and leave the chat context to see the result. This skill moves the same quiz into an IM-style conversation: users can answer in a thread, while the skill parses choices, asks for missing answers, scores them using the original web logic, and returns the personality result with a local image.

How It Works

The skill runs a fixed flow. It starts with a confirmation message, then sends 3 rounds covering 31 questions. Questions q1 through q30 are the formal set; drink_gate_q1 is fixed at the end of round 2, and drink_gate_q2 may be appended to round 3 when triggered. Each round sends its questions together. Users can answer with A/B/C/D, option text, or batched replies such as 1-a 2-c 3-b. The skill maps choices to numeric values, follows up on unclear or missing answers within the same round, and avoids carrying gaps into the next round. After all answers are collected, it calls a scoring script that returns JSON fields such as finalType, top3, dimensions, and imagePath. It then renders the result using the script output and sends the local image instead of exposing the raw path.

Boundaries

This is intended as an entertainment test, not a basis for psychological, medical, or life advice. Question wording, options, and personality descriptions should remain unmodified. Results depend on the fixed scoring implementation, including the hidden DRUNK path and the HHHH fallback. It fits lightweight chatbot interactions, not serious assessment.