OzLabs/ernest-nuclei-templates-v3

Ernest Nuclei模板数据集v3是一个用于训练模型从漏洞描述生成Nuclei安全扫描模板的综合数据集。该数据集包含11,590个训练示例，每个示例都包括唯一的标识符、结构化输入和JSON IR格式的输出。数据集分为CVE基础、CWE基础和通用类别，适用于代码生成模型的微调、少样本学习和结构化代码生成的研发。

The Ernest Nuclei Templates Dataset v3 is a comprehensive dataset for training models to generate Nuclei security scanning templates from vulnerability descriptions. It contains 11,590 training examples, each including a unique identifier, structured input, and JSON IR output. The dataset is categorized into CVE-based, CWE-based, and generic types, suitable for fine-tuning code generation models, few-shot learning, and research on structured code generation from specifications.