ai-security
SecurityLead story
Vercel acknowledges prompt-injection breach in AI code builder
A maliciously crafted project description caused the company's AI build agent to leak environment variables to an external URL.
Jaeden SchaferEditor in Chief6 min read