Learning Opportunities - A skill that helps with intentional skill development in Claude Code and Codex

It's a tool designed to address the 'Illusion of fluency' and 'Skill debt' problems that occur when users uncritically accept AI-generated code during Agentic Coding.

Designed for Claude Code and Codex, this skill suggests an optional 10-15 minute learning exercise whenever the user completes a major architectural task such as refactoring or schema changes.

Key features and characteristics:

• Applying learning science: It uses learning science techniques such as prediction, retrieval practice, and spaced repetition to run interactive exercises based on the user's actual project code.
• Various exercise modes: It offers various types that enhance metacognitive ability, such as 'Predict→Observe→Reflect', 'Debugging Prediction', and 'Explain the Codebase'.
• Workflow integration: It includes learning-opportunities-auto, which suggests learning at the time of a Git commit, and the orient skill, which helps users quickly grasp the structure of a new repository.

Beyond simply writing code on the user's behalf, it aims to maintain the user's expertise and improve their understanding of the codebase while collaborating with an AI agent.