CVE-2026-47285

The CVSS 6.5 score for this command injection vulnerability in VS Code is the most important analytical signal in the disclosure — and the most likely to be misread as reassuring. Command injection flaws routinely score critical because they provide direct arbitrary code execution. A medium rating demands explanation, and in VS Code's case, that explanation has real consequences for how you prioritize remediation. The vulnerability exists in a component capable of disclosing information over a network — the 'network' qualifier is doing significant work here. Most information disclosure CVEs are local: the attacker already has position and reads files they shouldn't. This phrasing suggests the compromised component can initiate outbound connections, making it an exfiltration channel rather than a passive disclosure. That's a qualitatively different risk profile. Three questions determine whether this merits priority treatment in your environment: First, identify the affected component. Is this in the main editor process, an extension host, the task system, or the integrated terminal? Each runs at different privilege levels. The extension host runs unsandboxed Node.js with filesystem and child_process access — if that's the vector, you're looking at a compromise of everything the extension API exposes. Second, what outbound connectivity does the affected component already have? VS Code connects to extension marketplaces, update servers, Git hosts, and cloud development environments. A command injection in a component that already trusts Microsoft's update infrastructure or GitHub APIs becomes a pivot to those trusted services — not just a local disclosure. Third, examine your deployment profile. Enterprise teams frequently pin VS Code versions, which means the disclosure-to-remediation gap may be weeks or months longer than typical. Each week of exposure adds new SSH keys, cloud sessions, and VPN connections to the machines running this vulnerable code. A 6.5 that looks acceptable at disclosure time compounds in exposure value. Until patch details clarify the component and its boundaries, treat this as a high-priority workstation vulnerability. The CVSS may be technically defensible under scoring rules that weren't designed for developer tooling — where a compromised workstation is a pivot point to every system the developer can reach, not an endpoint in itself.

Reviewed through automated stages and approved by a human before publication.