init-project-course

name	init-project-course
description	Use when you need to bootstrap a university course or module folder.
allowed-tools	Bash(mkdir*), Bash(mv*), Bash(ls*), Bash(tree*), Bash(find*), Bash(rm*), Bash(cp*), Read, Write, Edit, Glob, Grep, AskUserQuestion
argument-hint	[no arguments — runs in current directory]

Init Project Course

Bootstrap a university course or module folder into a clean, navigable structure. Designed for taught modules with lectures, workshops, assessments, and optionally recordings.

When to Use

• University modules with lectures + workshops + assessments
• Course folders that have accumulated files over a term
• When the user says "organise this module", "init course", "set up this module folder"
• Any taught course with recurring weekly components

When NOT to Use — Escalate

• To init-project-light if it's a one-off document collection (no lectures/workshops structure)
• To init-project-research if it's a research project that happens to be for a course

---

Phase 1: Scan

Read everything already in the directory before asking questions.

1. List all files and folders (excluding .claude/, .DS_Store)
2. Identify content types:
   • Lecture PDFs — files matching Lecture* or containing "lecture" in name
   • Seminar materials — files matching Seminar*, Session*, or containing "seminar" in name (discussion-based modules use seminars instead of lectures)
   • Lecture recordings — video files (.mp4, .mov, .mkv, .webm, .avi) or folders named recordings
   • Workshop folders — folders matching Workshop*, Lab*, Week*, or numbered folders with exercise content
   • Assessment folders — folders containing "group", "individual", "exam", "coursework", "assignment", "portfolio"
   • Workshop PDFs — PDFs inside workshop folders that are NOT lectures (exercise sheets, briefs)
   • Code projects — node_modules/, package.json, .py, .js, .ts files inside workshop folders
   • Notes — personal notes, reading lists, exercise solutions (common in student modules)
3. Detect duplicates: lecture PDFs that appear in multiple locations
4. Note workshop numbering gaps (normal — not every week has a workshop)
5. Check for existing organisation (already has lectures/, workshops/, etc.)

Goal: Build a complete inventory so the interview is short.

---

Phase 2: Interview (4-5 questions max)

Use the available structured-question mechanism. Only ask what you couldn't infer from Phase 1.

Pick from these (skip any you can already answer):

1. What module is this? — code + name (e.g., "IB9PK Advances in Behavioural Science")
2. Are you a student or instructor? — determines template variant (see Phase 3)
3. Who teaches it? — name(s) and role(s) (for student modules: instructor name; for instructor modules: module leader if different)
4. What components exist? — lectures/seminars, workshops, group work, individual work, exams, recordings
5. Status? — still in progress (more content coming) or archived/complete

If Phase 1 gave you enough, confirm your understanding instead of asking:

"This looks like [module code + name]. You're [student/instructor]. It has [N] workshops, [N] lecture PDFs, and [assessment types]. Still in progress. Correct?"

Always interview — even if scan is comprehensive. Course folders affect a full term of work; assumptions are expensive.

---

Phase 3: Organise

Present the proposed structure and wait for explicit approval before moving anything.

Full templates (student + instructor directory structures), workshop naming, lecture/recording/assessment handling rules: references/organise-templates.md

---

Phase 4: Create CLAUDE.md

Follow the lean-guidance-files rule. Include only:

1. Module overview — code, name, institution, programme, credits, 1-2 sentence description
2. People — instructor(s) and/or student, role
3. Assessment — type(s), format, word count/page limit, deadline(s) if known
4. Directory structure — compact tree of the organised layout
5. Conventions — detectable patterns (LaTeX compilation, GitHub repo naming, submission format)
6. Status note — "in progress" or "complete/archived", with note on what's still to come

Do NOT include:

• Lecture-by-lecture notes
• Workshop solution details
• Full assessment criteria (those live in docs/ or assessments/)
• Anything that duplicates global rules

---

Phase 5: Seed MEMORY.md

Create MEMORY.md using the appropriate template (student or instructor). Templates, settings.local.json, and vault entry details: references/memory-and-settings.md

---

Phase 6: Settings

Create .claude/settings.local.json if missing. Full details: references/memory-and-settings.md

---

Phase 7: Vault Sync

Offer to create entries in the vault tracking for this module (Student or Instructor). Database IDs and field mappings: references/memory-and-settings.md

---

Phase 8: Confirmation

Short report:

Set up course project: <module code + name>
Role: Student / Instructor

Created:
  - CLAUDE.md
  - MEMORY.md (seeded with [student/instructor] template)
  - lectures/ or seminars/ (N PDFs, deduplicated from M locations)
  - notes/ (student only, if applicable)
  - recordings/ (if applicable)
  - workshops/ (N workshops, renamed from original folders)
  - assessments/{types}/ (moved from original folders)
  - docs/ (module specification, guidelines)
  - [.claude/settings.local.json if created]
  - [vault entry created in Modules Pipeline (Student/Instructor)]

Workshop naming:
  "Workshop 1 - JavaScript Bootcamp" → workshops/01-javascript-bootcamp
  "Workshop 6" → workshops/06-cwe-vulnerability-analysis
  ...

---

Cross-References

Skill	Relationship
init-project-light	For non-course document collections
init-project-research	For research projects within a course
audit-project-course	Run later to check the structure is still clean
update-project-doc	Run later to refresh CLAUDE.md if the module grows