Polish generated textbook chapters while preserving structure, citations, and source grounding. Use when refining chapter-level Markdown, applying self-check reports, improving readability, standardizing terminology, enriching memory-oriented sections, and producing polished_index.json before PDF layout.
Parse files from manifest.json into clean structured text for downstream course-note agents. Use when extracting text from PDFs, PPTX, DOCX, HTML, Markdown, plain text, CSV, or image placeholders; cleaning noisy slide text with an LLM; detecting scanned/image-only materials; or producing parsed_index.json and parsed_documents/*.json.
Coordinate the full course_note_agents workflow from a local materials folder to reviewed Markdown and book-style PDF. Use when asked to run, resume, debug, or plan the whole multi-agent course-note pipeline, choose stage order, set run directories, or verify that outputs from ingestion, parsing, drafting, review, textbook writing, polishing, and layout connect correctly.
Generate source-grounded first-pass course notes from parsed course materials. Use when converting parsed_index.json into per-document Markdown/JSON drafts, extracting knowledge points, questions, answers, examples, memorization items, and open issues while preserving citations.
Convert polished course-note Markdown into LaTeX and book-style PDF. Use when rendering polished_index.json, checking CJK fonts, compiling with XeLaTeX, diagnosing LaTeX errors, or producing book.tex, book.pdf, compile.log, and layout_index.json.
Inventory and classify local course materials for course_note_agents. Use when scanning materials/, separating textbooks, lecture slides, assignments, past exams, solutions, images, notes, and miscellaneous files, checking duplicate files, or producing phase-1 manifest.json for later agents.
Audit course_note_agents pipeline outputs for correctness, coverage, and packaging readiness. Use when checking generated manifests, parsed text quality, review reports, textbook self-checks, polished Markdown, PDF compilation, config dry-runs, source leakage, API-key leakage, or deciding whether the workflow is reliable for a new course.
Review, correct, and enrich first-draft course notes. Use when checking draft notes against parsed sources, filling missing explanations, expanding terse content, separating questions from answers in mixed exam files, producing reviewed notes, and preserving unresolved issues instead of inventing facts.