English < Article List - DigitalToday

Search results for Adversa.AI

AI & Enterprise

AI coding tool could accelerate supply chain security threats

A vulnerability has been found that could allow supply chain attacks by abusing automation features in the AI coding tool Claude Code, SecurityWeek reported on May 7. Adversa.AI researchers said attackers could upload a GitHub repository containing hidden malware that Claude Code may automatically download. If a developer accepts a default trust prompt, malicious code could run with administrator privileges, enabling remote control. Risks rise in CI/CD pipelines, they said.

AI & Enterprise

Claude Code security flaw found days after source code leak

A critical security vulnerability has been found in Claude Code days after its source code was leaked. SecurityWeek reported the incident began when Anthropic distributed a debugging JavaScript source map to the npm repository during a Claude Code v2.1.88 update. Researcher Chaofan Shou (차오판 쇼우) shared the discovery on X, prompting analysis by developers worldwide. Adversa AI said it found a separate flaw that could allow bypassing command-blocking rules via prompt injection.