What problem does it solve? Learners often lose track of how their understanding of a concept changes over time, and AI assistants tend to write reflections for them instead of prompting genuine metacognition. This Skill guides a structured self-reflection session where the AI acts only as a mirror — quoting your past words and asking questions — while you write your own understanding into a persistent wiki page. ## Core Features & Use Cases - Guided reflection sessions: Prompts you to articulate your current understanding of a concept, source, or topic in your own words, then asks one targeted follow-up question (contrast, application, uncertainty, or implication). - Rewritable understanding with append-only history: Maintains a ## Current understanding section that always reflects your latest thinking, plus an ## Evolution log that records every session without modifying past entries. - Strict personal-layer isolation: Never edits academic wiki pages (concepts, sources, people) and never writes graph edges, keeping reflections as a private overlay on the knowledge base. - Use Case: After reading about cognitive offloading, say "reflect on cognitive-offloading" — the Skill quotes your previous entry, asks what has changed, and saves your new words to wiki/reflections/cognitive-offloading.md with an incremented evolution_count. ## Quick Start Tell the AI "reflect on <concept>" or invoke /lumi-learning-reflect, then answer its prompting questions in your own words so it can save your reflection page.
