Cloudforce One
@Cloudforce_One
The official threat intelligence account for @Cloudflare. Follow for threat research, incident assessments, WAF rule updates for emerging threats, and more.
4 Following 3.1K Followers
Cloudforce One conducted research into how linguistic deception and file structure can be used to bypass AI-driven code auditors across 18,400 API calls. The findings show that malicious detection rates drop when deceptive comments make up less than 1% of a file and that burying payloads in files larger than 3MB effectively blinds models to malicious intent. Read the full report here:
Show more
