Lugh

Local-first adaptive learning system. Any topic becomes a personalized audio course with built-in mastery verification. Named for the Celtic god of skill and craft.

Key Files

• Lugh — What it is, the 6-stage pipeline, design principles
• Welcome — Entry point and orientation
• The Pipeline — Full pipeline architecture (intake → episodes → assessment)
• The Gate — Mastery verification system
• Pre-Assessment — Initial knowledge check before course generation
• Episode Anatomy — Structure of a single learning episode
• Learning Depths — Depth levels (survey → working knowledge → mastery)
• Topic Organization and Pre-packaging — How topics are chunked and sequenced
• Listener Questions — Q&A mechanic for comprehension checks
• Shareables — Exportable artifacts from courses
• Shorts Mode — 1-minute video flashcard shorts for reviewing Gate-identified gaps
• Feynman Tutor Prompt — Feynman technique implementation
• The Akashic Records — Persistent learner model / memory layer
• Design Decisions — Architectural choices and rationale
• Known Limitations and Mitigations — What doesn’t work yet and workarounds

Walkthroughs

• Walkthrough Findings — Results from initial walkthroughs
• Walkthrough 2 - Understanding LLMs — Course on LLMs
• Walkthrough 3 - Understanding Socialism — Course on socialism

Courses

• Courses/ — Generated course artifacts

Design Principles

1. Local-first — no cloud dependency
2. Audio as primary medium
3. Mastery gating before progression
4. Feynman technique as verification

Maturity

Stage: Early experimentation. The pipeline is fully designed and a handful of manual walkthroughs have been completed (LLMs, Socialism). Those surfaced real findings that shaped the design. The automated pipeline — pre-assessment, content generation, gating, modality escalation — has not been built. The walkthroughs were handcrafted, not system-generated.