CVE-2026-81380
Copilot agent mode could send workspace data to an attacker
Microsoft · Visual Studio Code
First documented disclosure · Review checkpoint
Summary
Crafted instructions could cause the coding agent to send sensitive development data to an attacker service.
Severity and exploitation evidence
Medium · 5.3/10 · CVSS 3.1. Severity source.
Exploitation not established: Microsoft marks Exploited: No. The reviewed material does not establish malicious exploitation against users.
Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.
How the weakness works
Command injection redirects the agent’s handling of sensitive development context.
Conditions required
User interaction and environment-dependent conditions; Microsoft rates attack complexity high.
Potential impact
Exposure of workspace data and environment variables.
Affected versions and configurations
Visual Studio Code: 1.0.0 to before 1.136.2.
Fix and mitigation status
Visual Studio Code: 1.136.2.
Practical next steps
Install the vendor-listed corrected build for the applicable product and platform.
Operating lesson
Enforce the affected trust boundary independently of AI-generated content, caller identity and user-interface safeguards.
Evidence and limits
Historical backfill from a November 19–20 CVE update batch. Earliest reviewed public date is 2026-09-08; CNA initial publication is 2026-09-08, and current CNA datePublic is 2026-09-08T14: 00: 00.000Z. Microsoft CVSS 3.1 base score is 5.3; separate MSRC severity label: Important. The manufacturer marks Exploited: No in this reviewed entry. That is an attribution, not an assurance against future exploitation.
Sources
- Microsoft CNA authority record
Primary source ·
- MSRC original Sep CVRF entry
Primary source ·