course-review-from-materials

name	course-review-from-materials
description	Use when given course materials such as PPT, PDF, Markdown notes, or mixed lecture files and asked to produce a review outline, study plan, knowledge-point summary, selected high-frequency exam problems, formula sheets, or other exam-prep artifacts from them, especially when course context or likely test focus must be supplemented with web research.

Course Review From Materials

Overview

Turn heterogeneous course materials into a compact review pack that is usable for exam prep. Prefer structured Markdown outputs, keep claims tied to source materials, and separate verified knowledge from inference.

Workflow

1. Inspect the input set and classify each source by role.
2. Unless the user explicitly asks for local materials only, run web research to fill course context, standard topic naming, likely textbook structure, and common exam focus.
3. Build a coverage map for chapters, topics, formulas, algorithms, definitions, and examples.
4. Extract the exam-relevant core first: definitions, derivations, methods, templates, pitfalls, and representative problems.
5. Generate the default output set unless the user asks for a different shape.
6. Mark missing coverage, conflicting sources, inferred content, and externally researched context explicitly.

Required Web Research

Default rule: browse first, then write.

Skip web research only if the user clearly says the output must be based on local files only.

When browsing, prioritize:

• official course pages, syllabus pages, department pages, and instructor pages
• textbook tables of contents or publisher summaries
• authoritative teaching and learning resources for study design
• reputable course notes or archived exam-topic summaries when official material is thin

Use web research for:

• disambiguating course topic names
• reconstructing a more complete chapter map
• identifying common exam-worthy areas, recurring derivations, and standard problem types
• validating whether the local materials appear incomplete or unusually sparse

Do not let web research override explicit course-local rules from the provided materials. If they conflict, treat local course materials as authoritative for course-specific content and note the discrepancy.

Source Handling

• Treat PPT/PPTX as lecture-outline-first sources. They are useful for structure, keywords, and examples, but often omit proof steps and caveats.
• Treat PDF as high-fidelity course content. Use it for definitions, derivations, formulas, and authoritative wording when extraction is reliable.
• Treat MD and plain text notes as user-facing supplements. They often reveal emphasis, shortcuts, and likely exam focus, but may be incomplete.
• When the same topic appears in multiple sources, prefer the most complete and latest-looking version. Note conflicts instead of silently merging them.
• When extraction quality is poor, say so and continue with a best-effort summary rather than pretending certainty.

Default Outputs

Unless the user requests another format, produce these files in a dedicated output folder:

1. 00-source-map.md

   • Inventory of source files
   • Coverage by chapter/topic
   • External links consulted
   • Confidence notes, gaps, and assumptions

2. 01-review-outline.md

   • Chapter tree or topic tree
   • What must be memorized, understood, derived, or practiced
   • High-risk sections and dependencies between topics

3. 02-review-plan.md

   • Time-phased review schedule
   • Priority ranking by difficulty and exam importance
   • Short daily or session-based checkpoints

4. 03-knowledge-summary.md

   • Chapter-by-chapter knowledge points
   • Definitions, formulas, workflows, comparisons, common traps
   • Short memory cues where useful

5. 04-selected-exam-problems.md

   • Representative problems only, not a dump
   • Problem statement, tested concept, method outline, answer sketch
   • Include why the problem is high-value for revision

Add optional files only when they clearly help:

• 05-formula-sheet.md
• 06-active-recall-quiz.md
• 07-gap-and-risk-log.md
• 08-oral-recitation-outline.md

Use the templates in assets/templates/ as the default skeleton.

Output Rules

• Write in the user's language unless asked otherwise.
• Keep a clear boundary between source-backed, inferred, and missing.
• Label conclusions that come mainly from web research as externally supported.
• Prefer compact structure over long prose. Use tables when comparing concepts or methods.
• Do not fabricate page numbers, theorem names, or source quotations when extraction is uncertain.
• If materials are incomplete, still produce the pack and add a short "Missing coverage" section.
• If the course is computational, include solved examples and error-prone steps.
• If the course is theory-heavy, emphasize definitions, conditions, proof ideas, and theorem relationships.
• Make the review outline scan-friendly enough that a student can use it as a one-pass checklist the night before an exam.

Selecting Exam Problems

Choose problems that maximize revision value:

• canonical definitions or theorem applications
• standard derivations or algorithm traces
• common confusion points and near-miss variants
• past-paper style comprehensive questions if they can be reconstructed from the materials

For each selected problem, include:

• tested concept
• why it matters
• solution path
• final answer or result
• mistakes to avoid

If the exact original problem statement is unavailable, rewrite a faithful practice version and label it adapted from materials.

Review Design Patterns

Borrow proven study-material patterns where they help:

• use Cornell-style cue-and-summary compression for long notes
• add active-recall prompts instead of passive restatement
• group concepts with comparisons, dependencies, and pitfalls rather than flat bullet dumps
• use spaced-stage planning when the user has multiple days available
• include worked-example structure for high-value problems

Read references/review-design-principles.md when you need the design rationale.

Minimal Quality Bar

Before finishing, check:

• every promised output file exists
• chapter numbering is internally consistent
• formulas and notation are not silently changed across files
• the review plan references actual topics from the materials
• the selected problems cover multiple high-priority topics instead of clustering on one easy area
• web links consulted are listed when browsing was used
• the review outline is consistent with both local materials and any externally researched course context

Resources

• Read references/output-conventions.md when you need the detailed labeling scheme and file-by-file checklist.
• Read references/review-design-principles.md when you need guidance on study-guide structure, active recall, and outline design.
• Reuse the Markdown templates under assets/templates/ instead of improvising new structures for ordinary course-review tasks.