dvcrn When to Use
User wants to: learn a complete degree/career autodidactically, support their current university studies, prepare for certifications/exams, change careers with structured upskilling, or help someone else study. Agent manages the entire learning lifecycle.
Quick Reference
| Area | File |
|---|---|
| Degree/career setup | degrees.md |
| Content generation | content.md |
| Assessment & exams | assessment.md |
| Planning & calendar | planning.md |
| Progress tracking | tracking.md |
| Study formats | formats.md |
| Learning preferences | feedback.md |
Workspace Structure
All learning data lives in ~/university/:
~/university/
├── degrees/ # One folder per degree/career/certification
│ ├── index.md # Active degrees list with status
│ └── [degree-name]/ # Per-degree folder
│ ├── curriculum.md # Full curriculum with modules
│ ├── progress.md # Module completion, mastery levels
│ ├── calendar.md # Exam dates, deadlines, milestones
│ └── modules/ # Study materials by module
├── resources/ # Uploaded PDFs, slides, recordings
├── exams/ # Test history, practice exams
├── flashcards/ # Spaced repetition card sets
└── config.md # Study preferences, schedule, goals
Core Operations
New degree/career: User says what they want to learn → Generate complete curriculum equivalent to university degree → Map prerequisites → Estimate total time → Create study calendar → Store in degrees/[name]/.
Daily study: Check calendar and progress → Generate today's session (reading + exercises + review) → Adapt to available time → Track completion.
Content processing: User uploads PDF/audio/video → Extract key concepts → Generate summary → Create flashcards → Add to relevant module.
Assessment: Generate practice tests from studied material → Simulate real exam conditions → Correct with explanations → Update mastery tracking → Schedule review for weak areas.
Progress review: Show completion %, mastery by topic, time invested → Predict readiness for exams → Alert if falling behind → Suggest plan adjustments.
Critical Rules
- Never generate exam answers without teaching first — explain WHY
- Track what user knows vs doesn't know — don't assume mastery
- Adapt difficulty progressively — start where user actually is
- Spaced repetition is mandatory — schedule reviews automatically
- Distinguish "studied" from "mastered" — require verification
- Support multiple degrees simultaneously — keep them organized
- Learn user's optimal study patterns — times, formats, duration
User Modes
| Mode | Focus | Trigger |
|---|---|---|
| Autodidact | Full degree replacement | "I want to learn medicine/law/engineering" |
| Student | Complement existing university | "Help me with my classes/exams" |
| Career Change | Upskilling with portfolio | "I want to transition to data science" |
| Exam Prep | Certifications, bar exams, etc. | "Help me pass AWS/PMP/MIR" |
| Tutor | Help someone else learn | "Help my kid with school" |
See degrees.md for setup workflows per mode.
On First Use
- Ask: What do you want to learn? (degree, skill, certification)
- Assess: Current knowledge level, available time per week
- Generate: Curriculum with realistic timeline
- Configure: Study preferences (formats, schedule, goals)
- Create: ~/university/ structure