Moodle MCQ

A Claude Code skill for generating, reviewing, and improving Moodle quiz questions

Create exam-ready Moodle quizzes directly from Claude Code. Supports GIFT (~6x fewer tokens than XML), Moodle XML (with syntax highlighting), and Aiken formats. Enforces answer-length balancing, self-contained stems, and plausible distractors based on real pedagogical best practices.

---

Quick Start

# Install
mkdir -p ~/.claude/skills
git clone https://github.com/danielcregg/moodle-mcq.git ~/.claude/skills/moodle-mcq

# Use in Claude Code
/moodle-mcq
Generate 20 MCQ questions on Python data structures for first-year students

---

Four Modes

Mode	Invoke With	What It Does	Target Success Rate
Create Easy	/moodle-mcq easy	Revision quizzes, formative assessment, confidence building	85-95%
Create Standard	/moodle-mcq	Balanced conceptual questions with plausible distractors	75-85%
Create Challenging	/moodle-mcq challenging	Exam-grade questions testing judgment and trade-offs	60-75%
Review & Improve	/moodle-mcq review	Assess existing MCQs, fix weak distractors, rebalance lengths	N/A

---

Three Output Formats

Format	Tokens per 10 Qs	Best For	Moodle Import Path
GIFT (default)	~550	Standard quizzes, fastest generation	Question bank > Import > GIFT
Moodle XML	~3,500	Code with syntax highlighting, images, tags	Question bank > Import > Moodle XML
Aiken	~300	Quick review, simplest format	Question bank > Import > Aiken

GIFT uses approximately 6x fewer tokens than Moodle XML for the same questions. For a 50-question quiz, that is ~2,750 tokens (GIFT) vs ~17,500 tokens (XML).

---

Before & After

Create Mode: Easy vs Standard vs Challenging

Easy (85-95% success rate)	Standard (75-85% success rate)	Challenging (60-75% success rate)
Which keyword defines a class in Java? A) class <-- correct B) struct C) define D) object Tests: single-fact recall Distractors from other languages	What keyword declares inheritance in Java? A) extends <-- correct B) implements C) inherits D) super Tests: conceptual understanding Distractors are real Java keywords	Your AI generates a DB query using string concatenation. It passes all tests. Which issue has highest priority? A) SQL injection <-- correct B) SELECT * exposes PII C) Timing-safe comparison D) Missing rate limiting Tests: prioritization across 4 real security concerns

Review Mode: Before & After

Before (too easy, 90%+ success rate)

After (appropriate, ~60% success rate)

According to the lecture, what causes a workflow to begin in GitHub Actions? A) An event trigger defined in the YAML file B) A manual button click each time C) Scheduled cron job every hour D) Administrator approval Problems: References lecture B, C, D are obviously wrong Length imbalance

A developer wants their CI workflow to run when code is pushed. What causes a GitHub Actions workflow to begin execution? A) An event trigger defined in the workflow YAML that responds to push or pull request events B) A webhook configuration mapping Git operations to corresponding workflow files C) A GitHub App that monitors repository events and dispatches workflows through the Actions API D) A branch protection rule specifying which workflows must pass before merging Fixes: no lecture ref, all plausible, balanced lengths

---

Supported Question Types

Type	GIFT	XML	Aiken
Multiple choice (single answer)	Yes	Yes	Yes
Multiple choice (multi-answer)	Yes	Yes	-
True / False	Yes	Yes	-
Short answer	Yes	Yes	-
Numerical	Yes	Yes	-
Matching	Yes	Yes	-
Cloze / Fill-in-the-blank	Yes	Yes	-
Calculated	-	Yes	-
Drag and drop	-	Yes	-

---

Key Features

Answer Length Balancing (15/15/70 Rule)

Students exploit patterns when correct answers are systematically longer. This skill enforces a mandatory distribution and includes a verification table in every review document:

Correct Answer Position	Target	Purpose
Shortest option	~15%	Prevents "longest = correct" bias
Longest option	~15%	Prevents "shortest = correct" bias
Middle length	~70%	No exploitable pattern

How it works

1. After drafting all questions, the skill computes each option's character count (excluding HTML markup)
2. Classifies each correct answer as Shortest / Longest / Middle
3. Tallies across all questions to verify the 15/15/70 target
4. Rebalances if needed by adjusting distractor verbosity or trimming correct answers
5. Includes the distribution table in the review document for instructor verification

Self-Contained Stems

Every question must stand alone without external references:

Rule	Bad Example	Good Example
No lecture refs	"According to the lecture..."	"In Python..."
No slide refs	"As shown on slide 12..."	Context in the stem
No example refs	"In the Person class example..."	Include the actual code
Code included	"What does the setter do?"	Show the full code in the question

Five Distractor Strategies (Challenging Mode)

Click to expand

Strategy	Description	Example
Overcorrection	Takes a good practice too far	"Review every line with the entire team"
Outdated Practice	Was correct in older versions	"Use class components for all React code"
Wrong Context	Correct for different situations	"Use SELECT * for better performance"
Incomplete Solution	Addresses part of the problem	"Add unit tests" (when security review also needed)
Reasonable Misunderstanding	Based on incomplete knowledge	"Copilot adapts based on package.json versions"

Each distractor is tested against: Can an expert eliminate this without thinking? Would a B-grade student find it tempting? Does it represent a real misconception?

Code Syntax Highlighting (XML Mode)

When generating Moodle XML, code questions include inline CSS syntax highlighting that renders correctly in Moodle without any plugins:

Java: blue keywords, green numbers, red strings, green comments Python: blue keywords, teal built-ins, green numbers, red strings

Example: Java code in Moodle XML

<questiontext format="html">
  <text><![CDATA[<p>What happens when a username longer than 10 characters is passed?</p>
<pre style="background-color: #f4f4f4; padding: 10px; border-radius: 5px;
  border-left: 3px solid #4CAF50; font-family: 'Courier New', monospace;">
<span style="color: #0000ff;">public void</span> setUsername(String username) {
    <span style="color: #0000ff;">if</span> (username.length() > <span style="color: #098658;">10</span>) {
        <span style="color: #0000ff;">this</span>.username = username.substring(<span style="color: #098658;">0</span>, <span style="color: #098658;">10</span>);
    }
}</pre>]]></text>
</questiontext>

Moodle XML Compatibility

Built-in fixes for common Moodle import issues:

Issue	Fix
<n> instead of <name>	Enforces full <name> tag (prevents "Missing question name" error)
Penalty tags	Omitted entirely (no penalty for wrong answers)
Generic titles	Requires descriptive names (e.g., "Array_Index_Exception")
Missing CDATA	All text wrapped in <![CDATA[...]]> sections
Category format	Uses $course$/Category Name structure

---

Review Document

Every generation includes a human-readable review document (_review.md) for instructor verification:

# Python Data Structures - Quiz Review

## Difficulty Mode: Standard
## Output Format: GIFT

## Length Distribution Summary
| Correct Answer Position | Count | Percentage | Target |
|------------------------|-------|------------|--------|
| Shortest               | 3     | 15%        | ~15%   |
| Longest                | 3     | 15%        | ~15%   |
| Middle                 | 14    | 70%        | ~70%   |

## Category: Lists and Tuples

### Question 1: List_vs_Tuple_Mutability
*Correct answer length: MIDDLE*

What is the key difference between a list and a tuple in Python?

A) Tuples are faster but have no other difference
B) **Lists are mutable, tuples are immutable**
C) Lists can store multiple types, tuples cannot
D) Lists use square brackets, tuples use parentheses

---

GIFT Format Example

$CATEGORY: Programming/Python/Data Types

::List vs Tuple::What is the key difference between
a list and a tuple in Python?{
=Lists are mutable, tuples are immutable#Correct!
  Lists can be modified after creation while tuples cannot.
~Lists use square brackets, tuples use parentheses#Syntax
  differs, but the key functional difference is mutability.
~Lists can store multiple types, tuples cannot#Both
  can store mixed data types.
~Tuples are faster but have no other difference#Tuples
  are slightly faster, but immutability is the key difference.
}

Note for programming questions: GIFT requires escaping ~ = { } # : characters. For code-heavy questions, Moodle XML with <![CDATA[<pre>...</pre>]]> is recommended to avoid escaping issues.

---

File Structure

~/.claude/skills/moodle-mcq/
  SKILL.md        # Skill definition (the brain)
  README.md       # This file
  LICENSE          # MIT

---

Compatibility

Platform	Supported	How
Claude Code (CLI)	Yes	Install as skill, invoke with /moodle-mcq
Claude Desktop	Yes	Add as skill via Settings > Skills
Moodle 4.x	Yes	Import generated GIFT/XML/Aiken files
Moodle 3.x	Yes	Import generated GIFT/XML/Aiken files
Windows / macOS / Linux	Yes	All platforms supported

---

Installation

Claude Code (CLI)

mkdir -p ~/.claude/skills
git clone https://github.com/danielcregg/moodle-mcq.git ~/.claude/skills/moodle-mcq

Verify: type /moodle-mcq in Claude Code — if the skill loads, you are good to go.

Claude Desktop

1. Download SKILL.md from this repo (or download the repo as a .zip)
2. Open Claude Desktop
3. Go to Settings > Skills
4. Click Upload skill
5. Drag and drop the SKILL.md file, or the .zip file containing it

File requirements:

• .md file must contain skill name and description formatted in YAML frontmatter
• .zip or .skill file must include a SKILL.md file

Manual Download

Download SKILL.md from this repo and place it at:

~/.claude/skills/moodle-mcq/SKILL.md

---

Usage Examples

Generate easy revision questions

/moodle-mcq easy

Generate 10 easy revision questions on Python basics (variables, loops, functions)
for students preparing for their first exam.

Generate a standard quiz from lecture slides

/moodle-mcq

I have a lecture on Java OOP (inheritance, polymorphism, encapsulation).
Generate 15 MCQ questions in GIFT format, 5 per topic.

Generate a challenging exam

/moodle-mcq challenging

Create 20 hard questions on Python error handling and exceptions.
Use Moodle XML format with syntax-highlighted code snippets.
Target 60-70% success rate.

Review and improve existing questions

/moodle-mcq review

[paste existing questions in any format]

Fix weak distractors, balance answer lengths, remove any lecture references.

Generate from PowerPoint slides

/moodle-mcq

Extract content from lecture.pptx and generate 25 questions across all topics.

The skill uses python -m markitdown lecture.pptx to extract slide content when PowerPoint files are provided.

---

Importing into Moodle

1. Log into Moodle as a teacher or admin
2. Navigate to your course
3. Go to Question bank > Import
4. Select the format matching your file:
   • .gift files: select GIFT format
   • .xml files: select Moodle XML format
   • .txt files: select Aiken format
5. Upload the file and click Import
6. Review the imported questions in the question bank

---

Version History

Version	Date	Changes
3.0.0	2026-03-24	Major merge: combined 4 skills into one. Added GIFT + Aiken formats. Added Python syntax highlighting. Published to GitHub.
3.1.0	2026-03-24	Added Easy difficulty mode (85-95% success rate) for revision quizzes and formative assessment. Four modes: easy, standard, challenging, review.
2.3	2026-02-12	Merged reviewer skills into single mcq-reviewer
2.2	2026-02-12	Merged creator skills with difficulty mode parameter
2.1	2025-12-03	XML tag docs fix, length distribution tables
2.0	2025-11-06	Difficulty calibration overhaul, 5 distractor strategies
1.4	2025-10-30	Java syntax highlighting CSS template
1.3	2025-10-30	Self-contained code requirement
1.2	2025-10-30	Critical <name> tag fix for Moodle import
1.0	2025-10-23	Initial skill for AI Assisted Programming module

---

Contributing

Contributions are welcome. If you have ideas for improving question generation quality, supporting additional Moodle question types, or fixing edge cases:

1. Fork this repository
2. Create a feature branch (git checkout -b feature/your-feature)
3. Commit your changes
4. Push to the branch (git push origin feature/your-feature)
5. Open a Pull Request

---

License

MIT - use it however you like.