unified-study-note

Synthesize slides, transcripts, code, and quizzes into an Obsidian master study note.

Updated May 10, 2026
One-click install
npx skills add https://github.com/cookkie03/skills --skill unified-study-note-cookkie03
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: unified-study-note
Source: https://github.com/cookkie03/skills/tree/main/unified-study-note
Command: npx skills add https://github.com/cookkie03/skills --skill unified-study-note-cookkie03

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pypdf, pymupdf, requests, and includes scripts (resource) components.

What problem does it solve? University course materials are scattered across slide decks, audio recordings, code notebooks, workbooks, and quizzes, making exam preparation fragmented and incomplete. This Skill consolidates every source into a single, exhaustive, concept-organized Obsidian master note without losing any detail. ## Core Features & Use Cases - Deterministic Source Diffing: Scans course folders recursively and uses SHA-256 hashes to detect only new or modified files, avoiding redundant reprocessing. - Multi-Format Ingestion: Extracts slide text and renders slide images from PDFs, transcribes lecture audio via OmniRoute STT with automatic chunking, and reads code notebooks and markdown workbooks directly. - Map-Reduce Synthesis Pipeline: Classifies source lines into topics with an LLM classifier, builds zero-token topic payloads deterministically, and dispatches sub-agents for lossless topic-by-topic writing into the master note. - Audit & Verification: Rebuilds a complete source record manifest and validates TOC links, code fences, and math delimiters. - Use Case: A data science student with a semester of PDF lectures, recorded classes, R scripts, and quizzes runs the pipeline to produce one definitive, exam-ready Obsidian study guide with citations, formulas, and embedded diagrams. ## Quick Start Compile all materials in my course folder into a unified Obsidian master study note for this course.

Frequently Asked Questions about unified-study-note

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I combine lecture slides and audio recordings into one study note?

Run the pipeline on your course folder: it extracts slide text and images from PDFs, transcribes audio via OmniRoute STT, classifies all content by topic, and synthesizes everything into a single Obsidian master note with citations.

How does the skill avoid reprocessing already-ingested course files?

The diff_sources.py script compares filenames and SHA-256 hashes against the source record manifest in the master note. If no new or modified sources are found, it exits immediately without spending any LLM tokens.

What file formats are supported for course material ingestion?

Supported formats include PDF and PPTX slides, m4a/mp3/wav/aac/webm/ogg audio, R/Python/SQL/ipynb code files, and docx/txt documents. Markdown workbooks and personal notes are read directly in place.

Does the transcription handle long lecture recordings?

Yes, transcribe_audio.py automatically slices audio longer than the chunk threshold into 10-minute MP3 segments using ffmpeg, transcribes each via OmniRoute STT, and concatenates the results into one transcript file.

Why does slide image rendering fail on my machine?

Rendering requires one of three backends: macOS native Swift/PDFKit, pdftoppm from poppler, or PyMuPDF. If none are installed, the script skips image rendering with a warning but still extracts slide text.

What are the limitations of the LLM topic classification step?

The classifier requires a reachable OmniRoute-compatible endpoint and processes files in 500-line chunks to avoid context overflow. Misaligned line indexes from the LLM are clamped safely by the payload builder, but classification quality depends on the chosen model.