Personal Tutor for Junior High

Problem

Junior-high tutoring often breaks down in three ways: students can solve a problem but not explain it, materials are scattered, and repeated practice tends to drift away from the textbook, grade level, or error log. This skill puts core junior-high subjects—Chinese, math, English, physics, chemistry, biology, history, geography, and civics—into a Q&A assistant aimed at homework, preview, review, and exam prep rather than generic chat.

How It Works

The skill invokes a local open-source tutoring engine, with the agent selecting the appropriate command when a learning intent appears. Capabilities include:
- chat: answer questions with an attached knowledge base or normal context;
- deep_solve: solve a single problem step by step with citations and validation;
- deep_question: generate practice questions by chapter, concept, or difficulty, with answers and explanations;
- deep_research: run parallel multi-agent research and produce cited study material;
- visualize: render concepts or algorithms into interactive pages for abstraction-heavy topics;
- mastery_path: inspect a mastery path and assessment status.

A typical flow is to import textbooks, exams, and mistake notes into a knowledge base, then run tasks such as “this chapter is unclear,” “drill this concept,” or “review this exam topic.” The closer the materials are to the student’s actual level, the easier it is to keep outputs within junior-high scope.

Boundaries

It suits students in grades 7–9, parents supporting study, tutors, and teachers preparing lessons; it is not intended for open-ended consultation without local materials. It requires a local runtime, a configured LLM API key, and stores the knowledge base and settings locally by default. Sensitive exams or mistake logs should be handled in a trusted environment.