Summary Lineage Skill transforms dense learning materials like courses, books, and long documents into source-backed "Agent Skills," creating reusable, actionable knowledge systems. It targets users and developers aiming to leverage AI agents for learning, decision-making, and work output, converting passive content into an active, transferable method system.

How It Works The system uses a "distillation" pipeline (Capture, Cite, Compress, Connect, Codify, Evaluate) to process diverse inputs. It uniquely extracts "capability assets"—diagnostics, workflows, rubrics, templates, transfer rules, and failure modes—beyond simple summaries. This codifies expert methods into agent-callable skills, enabling AI to act as a mentor, consultant, or practitioner with traceable sources.

Quick Start & Requirements Installation is via agent command referencing docs/install.md. Prerequisites include input materials (video, audio, PDF, text), local tools (bash, git, python3, pip, ffmpeg), and configured OpenAI-compatible APIs for transcription, vision, and text processing. API keys and model endpoints are essential; an optional MinerU token is available for advanced PDF parsing.

Highlighted Details

• Extracts structured "capability assets" (diagnostics, workflows, rubrics, templates, transfer rules, failure modes).
• Generates OKF-compatible knowledge packages (Markdown + frontmatter) for progressive directories.
• Features enhanced source traceability via scripts like fetch_course_evidence.py.
• Supports multi-course merging while preserving source boundaries.
• Core pipeline: Capture, Cite, Compress, Connect, Codify, Evaluate.
• Example: nihaisha-tcm project distilled 100GB+ video material.

Maintenance & Community Acknowledges Datawhale, Ma Laoshi (LSGOGROUP), rfeng1016, and LINUX DO community for contributions and inspiration. Direct community links (Discord/Slack) are not specified.

Licensing & Compatibility Licensed under PolyForm Noncommercial License 1.0.0. Commercial use is prohibited and requires direct contact with the maintainer (juneyaooo@gmail.com).

Limitations & Caveats The noncommercial license restricts business adoption. Effectiveness depends on input material quality and configured external LLM/vision APIs, which may incur costs. The technology is advanced and likely evolving.