数据集开放下载：带CVE标签的真实HTTP流量

DataCon 2025 publishes a CVE-labeled HTTP traffic dataset for a vulnerability attack identification competition, providing train/test JSON Lines gzip files containing zlib+Base64-encoded HTTP request payloads, label metadata, and about 1000 CVE identifiers. The document details data collection from honeypots (July 2024–July 2025), automated CVE annotation and known label error rates, de-duplication, sampling and split methodology, competition scoring rules, and recommended external resources for participants.