CVE-2026-55145

Outlook Copilot could process untrusted commands as trusted content

Microsoft · Microsoft Copilot

First documented disclosure · Review checkpoint

Summary

An authorized attacker could influence Outlook Copilot through command injection.

Severity and exploitation evidence

Medium · 6.3/10 · CVSS 3.1. Severity source.

Exploitation not established: Microsoft marks Exploited: No. The reviewed material does not establish malicious exploitation against users.

Exploitation evidence source

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

Untrusted content reaches Copilot without the intended isolation.

Conditions required

An attacker with the vendor’s low-privilege prerequisite supplies content that a user causes Copilot to process.

Potential impact

Tampering, with confidentiality impact also reflected in the vendor CVSS assessment.

Affected versions and configurations

Microsoft Copilot: affected service; numeric range not published.

Fix and mitigation status

Microsoft supplies an isolation mitigation contingent on Outlook external-sender tagging; no numeric fixed build is listed.

Practical next steps

Administrators must enable Outlook external-sender tagging through Exchange Online PowerShell so Microsoft’s mitigation can isolate untrusted content before it reaches Copilot.

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 2026-07-14; CNA initial publication is 2026-07-14, and current CNA datePublic is 2026-07-14T14: 00: 00.000Z. Microsoft CVSS 3.1 base score is 6.3; separate MSRC severity label: Moderate. The manufacturer marks Exploited: No in this reviewed entry. That is an attribution, not an assurance against future exploitation. This entry does not carry the no-customer-action cloud FAQ. The CVSS temporal exploit-maturity field is Proof of Concept (E:P); Microsoft still marks Exploited: No.

Sources