I Am a Top Student AI-Assisted Learning System

Problem It Addresses

Children's self-study often gets stuck on practical issues: not knowing what to review today, getting generic AI answers instead of subject-specific tutoring, and lacking visible progress for parents or teachers. This skill turns those issues into a maintainable local learning system: a browser dashboard tracks progress and review tasks, a Flask data API stores mastered and error records, and subject teacher prompts constrain how the AI responds.

How It Works

• Local dashboard: dashboard.html provides progress tracking, review tasks, and topic graphs.
• Data layer: data_api.py uses a Flask backend for read/write operations and exposes a local service.
• Teacher prompts: teachers/*.md defines subject personas and rules instead of leaving the model unconstrained.
• Review scheduling: generates Ebbinghaus review plans using the 1-2-4-7-15-30 day cycle.
• Adding a subject: copy teacher-prompt-template.md, fill in the 9 required modules, create records/<subject>/mastered.md and errors.md, then register the subject in the dashboard.

Each subject needs a positioning type, such as skip-ahead, extension, written-content-focused, or interest-driven. Global constraints also prevent out-of-scope teaching, for example no pronunciation or speaking for English and no pre-textbook formulas for Math.

Boundaries

• Best suited to desktop setups with a local Python environment; the core experience depends on a browser and local service.
• Prompt quality matters more than generic model tuning; edit teachers/*.md first.
• This is not a textbook; assets/teachers/*.md mainly provides examples and constraints.