guychuk/ernest-nuclei-templates-v1-instruct

Ernest Nuclei Templates Dataset v1 (Instruction Format)

This is a restructured version of the Ernest Nuclei Templates dataset optimized for instruction tuning.

Format

Each example has this structure:

{
  "instruction": "Generate a Nuclei template for CVE-2024-1234",
  "input": "Title: XSS in Product\nSummary: Description...\nSeverity: high",
  "output": "id: cve-2024-1234\n\ninfo:\n  name: ...\n\nhttp:\n  - method: GET\n    ...",
  "category": "cve",
  "severity": "high",
  "id": "CVE-2024-1234"
}

Usage

from datasets import load_dataset

dataset = load_dataset('json', data_files={
    'train': 'train.jsonl',
    'test': 'test.jsonl',
    'validation': 'validation.jsonl'
})

# For instruction tuning
example = dataset['train'][0]
print(f"Instruction: {example['instruction']}")
print(f"Input: {example['input']}")
print(f"Output: {example['output']}")

Advantages

• Flatter structure: Easier to work with in ML frameworks
• Instruction format: Ready for instruction-tuning (Alpaca, GPT-4 style)
• Text-focused: Input and output are plain text strings
• Smaller size: ~30% smaller than nested format
• Faster loading: Less parsing overhead

Original Dataset

See the original nested format at: ../ernest-nuclei-templates-v1/