Study Coach — Persona

The problem

Standard LLMs tend to provide immediate answers, which fosters student dependency and bypasses the actual learning process. Developers building educational tools struggle to prevent AI from simply completing homework for users rather than teaching them.

What it does

• Uses Socratic questioning to guide students toward answers through their own reasoning.
• Implements evidence-based study techniques like spaced retrieval and active recall.
• Provides progressive hints in "ladders" to ensure students struggle productively before receiving help.
• Enforces strict boundaries against completing graded work, essays, or exam problems.
• Requires students to explain concepts back in their own words to verify comprehension.

Why this beats prompting it yourself

Direct prompting often fails to maintain a consistent educational persona, eventually collapsing into "here is the answer" when a user pushes back. This skill uses a sophisticated behavioral framework that prioritizes retrieval practice over simple response generation, ensuring the agent remains a coach rather than a ghostwriter.

Use cases

• Deploy a persistent tutor for test prep apps that refuses to do the work.
• Build a study companion that generates spaced repetition schedules based on notes.
• Create an interactive coding tutor that debugs via questioning instead of code blocks.
• Implement a safe learning environment for K-12 students with age-appropriate filters.

Known limitations

The coach will not grade official work or predict exam scores. It explicitly refuses to write essays or solve take-home exam problems.