CVE Chat-Style Multi-Turn Cybersecurity Records from 1999 to 2025

Trendyol's dataset contains approximately 300,000 Common Vulnerabilities and Exposures (CVE) records published between 1999 and 2025. Each record has been parsed, enriched, and converted into a conversational format. The dataset is hosted on Hugging Face and was last updated in June 2025.

Use Cases

• Train cybersecurity chatbots based on the conversational format of CVE records.
• Evaluate AI-Agent systems for vulnerability analysis using the multi-turn dialogue structure.
• Fine-tune large language models on domain-specific cybersecurity knowledge from the enriched CVE data.
• Generate synthetic cybersecurity Q&A pairs for testing based on the parsed vulnerability information.

Strengths

• Approximately 300,000 CVE records, described as the largest publicly available dataset of its kind.
• Covers a 26-year time range from 1999 to 2025, providing historical vulnerability context.
• Data is parsed, enriched, and formatted specifically for conversational AI applications.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is an approximation; exact size and file formats are unknown.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Trendyol, via Hugging Face.

Collection Method

CVE records were parsed, enriched, and converted into a conversational format.

Time Range

1999 to 2025

Freshness

Last updated 2025-06-12 12:15:23.

Geography

Global (CVE identifiers are international).