Researchers have identified two malicious Visual Studio Code extensions posing as AI-powered coding assistants. These extensions, which have a combined total of 1.5 million installs, are still available on the Visual Studio Marketplace. The first, "ChatGPT - 中文版," has over 1.3 million installs, while "ChatGPT - ChatMoss" has about 151,000. Both extensions secretly capture every file opened and any source code changes, sending this data to servers in China. The spyware operates without user consent, and the malicious code is hidden beneath legitimate functionality, such as autocomplete suggestions and error explanations. Koi Security, which uncovered these extensions, has dubbed the campaign MaliciousCorgi. The spyware not only exfiltrates file content but also includes a real-time monitoring feature that can trigger up to 50 file exfiltrations remotely. A hidden iframe in the extensions collects data through four Chinese analytics SDKs: Zhuge.io, GrowingIO, TalkingData, and Baidu Analytics. In a related security issue, six zero-day vulnerabilities have been found in JavaScript package managers like npm and pnpm. These flaws, called PackageGate, allow attackers to bypass security measures that prevent automatic execution of lifecycle scripts during package installations. While pnpm, vlt, and Bun have patched their vulnerabilities, npm has chosen not to address them, shifting the responsibility to users for vetting packages. GitHub, the parent company of npm, is actively looking for malware in the registry but emphasizes user trust in package repositories. Security experts recommend disabling scripts and carefully managing lockfiles, but caution that these measures alone are insufficient until the vulnerabilities are fully resolved.