Learning Workflow

This document describes the end-to-end workflow of how learning is organized and executed in SkillPilot, from static data to individual mastery.

1. Data Foundation: Curricula & Modules

The world of SkillPilot is built from Curricula (e.g., "Gymnasiale Oberstufe Hessen") and Modules (e.g., "Math Q1", "Physics Mechanics"). - Storage: These are stored as JSON files in the curricula/ folder. - Structure: They form a directed acyclic graph (DAG) of learning goals connected by requires edges.

2. The Learning Lifecycle

The learning process follows a strict sequence of personalization steps:

Step 1: Base Curriculum Selection (Level 1)

Before learning starts, a Base Curriculum must be chosen. - Definition: The complete set of all possible modules and goals defined by an authority. - Action: The user selects "Gymnasiale Oberstufe" or "B.Sc. Physics".

Step 2: Personal Curriculum (Level 2)

The learner selects the specific Modules relevant to their path. - Definition: A subset of the Base Curriculum (e.g., specific electives or majors). - Action: The learner chooses "Math (Advanced)", "Physics (Basic)", and omits "Latin". - Result: This defines the personal "search space" for the frontier calculation.

Step 3: Concrete Learning Goal (Level 3)

The learner sets a specific focus. - Definition: A target goal or topic to work towards. - Constraint: Only one Planned Goal can be active at a time to ensure clear focus. Setting a new goal replaces the previous one. - Default: "All goals in the Personal Curriculum" (if no specific goal is planned). - Action: The learner says "I want to learn Analysis" or "I want to finish my Bachelor's".

Step 4: The Frontier Loop (AI Assisted)

Learning happens along the Frontier. - The Frontier: The set of goals where: 1. The goal itself is not yet mastered. 2. All direct requires (prerequisites) are mastered. - Process: 1. Calculate Frontier: The system analyzes the graph and current mastery to find the "next best steps". 2. AI Guidance: An AI Tutor (ChatGPT) uses this frontier to suggest topics, explain concepts, and provide exercises. 3. Reverse Traversal: If a user wants a goal not on the frontier, the system traces back the requires chain to find the missing foundations.

Step 5: Mastery & Feedback (Level 4)

Success is recorded as Mastery. - Action: When a learner solves tasks correctly, the AI updates the mastery level (0.0 to 1.0) for that specific goal UUID. - Effect: - Mastering a goal satisfies the prerequisites for other goals. - The Frontier moves forward, opening up new learning opportunities. - The cycle repeats.