Quick Start

Bash
# First, read the assignment files
cat "grader-value/grading-ftds-p0/${FILE_QUESTION}"
cat "grader-value/grading-ftds-p0/${FILE_GRADING}"

# Check existing progress
ls "multiple-git-clone/${OUTPUT_LOCATION}/"
cat "multiple-git-clone/${OUTPUT_LOCATION}/todo.md" 2>/dev/null || echo "File belum ada"

Workflow

Progress:

• Read assignment question and grading rubric
• Initialize/read todo.md and review.md files
• Identify next student to grade
• Evaluate student submission against rubric
• Create grading.md file for student
• Update todo.md with progress and scores
• Update review.md with summary
• Prompt for next student

1. Initialize Files

Bash
# Create todo.md if not exists
echo "# Progress Penilaian ${CODENAME}

• Belum dimulai

• Belum ada data

• Belum ada data

Create review.md if not exists

echo "# Review Penilaian ${CODENAME}

" > "multiple-git-clone/${OUTPUT_LOCATION}/review.md"

### 2. Grade Each Student
For each folder in `multiple-git-clone/${OUTPUT_LOCATION}/`:
- Skip if `grading.md` already exists
- Analyze all files (except readme.md and assignment-rubrics.png)
- Check for AI usage indicators
- Score against each rubric point

### 3. AI Detection Criteria
- Overly polished writing for junior level
- Perfect formatting without errors
- Complex explanations beyond expected level
- Generic/template-like responses
- Inconsistent coding style

# Examples

**Example 1: Student Folder Structure**

multiple-git-clone/output-GC2-S2/ ├── student_001/ │ ├── analysis.py │ ├── report.md │ └── data/ ├── student_002/ │ ├── grading.md # Skip this one │ └── solution.ipynb └── todo.md

**Example 2: Grading Workflow**
```markdown
# Checking student_001
Files found: analysis.py, report.md, data/sales.csv

• Code style: Consistent, beginner-appropriate
• Explanations: Match junior level
• Probability: Rendah

Point 1 (Data Loading): 8/10 Point 2 (Analysis): 6/10 ...

# Best Practices

• Poin Penuh: Meets all requirements perfectly
• Poin Sebagian: Meets some requirements, has minor issues
• Tidak Ada Poin: Missing or fundamentally incorrect

• Be specific and actionable
• Highlight both strengths and improvements
• Use encouraging but honest language
• Focus on learning opportunities

• Look for complexity mismatches
• Check code commenting patterns
• Evaluate explanation sophistication
• Consider submission timing patterns

• Update todo.md after each student
• Calculate running averages
• Document common issues
• Maintain clear status

Common Pitfalls

1. Inconsistent Grading: Don't change standards between students
2. Vague Feedback: Avoid "good job" without specifics
3. Skipping Files: Always check for existing grading.md
4. Missing Updates: Update both todo.md and review.md
5. Language Mixing: Keep all feedback in Indonesian
6. Over-penalizing: Remember these are junior analysts
7. Ignoring AI Signs: Always assess AI usage probability

• Never overwrite existing grading.md files
• Always backup todo.md before major updates
• Check file permissions before writing
• Validate markdown formatting

• Use same criteria across all students
• Document reasoning for partial points
• Be fair but maintain standards
• Consider effort vs. results

Template Usage

Use the exact markdown templates provided for:

• grading.md - Detailed student feedback
• review.md entry format
• todo.md progress tracking

Always include:

• Specific code references
• Clear improvement suggestions
• Positive reinforcement
• Grade justification
• AI assessment reasoning

Remember: You're helping junior analysts grow, not just scoring them.