MEDIUM 4.4

CVE-2026-57966: Path Traversal in spice-vdagent File Transfers

A path traversal flaw in spice-vdagent allows a malicious or compromised SPICE host to write arbitrary files anywhere on a guest system. When a SPICE host initiates a file transfer, the agent doesn't validate the filename, enabling an attacker to place files in sensitive directories with the privileges of the logged-in user. This is a real but constrained risk: it requires either control of or a compromise of the SPICE host itself, not just network access.

Source data · NVD / CISA · public domain

CVSS
3.1 · 4.4 MEDIUM · CVSS:3.1/AV:L/AC:L/PR:H/UI:N/S:U/C:N/I:H/A:N
Weaknesses (CWE)
CWE-22
Affected products
6 configuration(s)
Published / Modified
2026-06-29 / 2026-07-08

NVD description (verbatim)

A path traversal vulnerability was found in spice-vdagent. This flaw allows a malicious or compromised SPICE host to write arbitrary files to any location on the guest operating system. This occurs because the filename provided by the SPICE host during file transfers is not properly sanitized before being used. An attacker could exploit this to write to sensitive locations with the privileges of the spice-vdagent process, typically the logged-in user. This issue requires the SPICE host to be untrusted or compromised for exploitation.

2 reference(s) · View on NVD →

SEC.co analysis · AI-assisted, reviewed against source

Technical summary

CVE-2026-57966 is a CWE-22 path traversal vulnerability in spice-vdagent's file transfer handling. The vulnerability stems from insufficient input sanitization on filenames received from a SPICE host during file operations. An attacker positioned as or controlling the SPICE host can craft filenames containing directory traversal sequences (e.g., '../') to write files outside intended directories. The write operation executes with the privilege level of the spice-vdagent process—typically that of the logged-in user—meaning sensitive application or system files could be overwritten or created. The attack surface is local to the guest environment; remote exploitation is not possible without prior compromise of the SPICE infrastructure.

Business impact

In virtualized or cloud environments leveraging SPICE for guest management, this vulnerability creates a post-compromise escalation path. If a SPICE host is breached or if an insider has malicious intent toward guest systems, attackers gain the ability to plant malware, modify configuration files, inject code into startup sequences, or disable security tools—all running as a regular user without requiring additional privilege escalation. For organizations running untrusted or semi-trusted SPICE hosts, the risk is elevated. For environments with strict host integrity controls, impact is substantially reduced.

Affected systems

Red Hat Enterprise Linux systems using spice-vdagent are affected across multiple versions. The spice-vdagent package itself is the direct vector. Any guest operating system running a vulnerable version of spice-vdagent on a SPICE-capable hypervisor (QEMU/KVM with SPICE protocol) is in scope. Windows and Linux guests are both potentially affected if running the vulnerable agent. Verify the exact affected versions against Red Hat security advisories and the upstream SPICE project.

Exploitability

This vulnerability requires an attacker to already control or compromise the SPICE host—a privileged position in a virtualization infrastructure. It does not allow unauthenticated remote code execution over the network. Exploitation is straightforward once that precondition is met: craft a file transfer request with a traversal payload in the filename. The CVSS score of 4.4 (MEDIUM) reflects this requirement; the local attack vector and high integrity impact are balanced by the need for high privilege on the host side. No public exploit code or active weaponization is currently tracked.

Remediation

Apply vendor security updates to spice-vdagent as released by Red Hat and the SPICE project. Patches will implement proper filename sanitization, stripping or rejecting directory traversal sequences before file write operations. Verify patch application across all guest systems in your environment. As an interim control, restrict SPICE host access to trusted administrators and monitor file transfer activity in guest logs.

Patch guidance

Consult Red Hat's security advisory for CVE-2026-57966 to identify the patched spice-vdagent version for your Enterprise Linux release. Test patches in a non-production environment first. Patch deployment should prioritize systems in high-trust environments or those hosting sensitive workloads. Coordinate patching with your virtualization team to ensure guest consistency and to validate that file transfer functionality is not disrupted post-patch. Consider using configuration management tools to standardize versions across your fleet.

Detection guidance

Monitor spice-vdagent process activity and file write operations for suspicious patterns: files being written to system directories, unexpected modifications to sensitive paths, or filenames containing '../' or absolute paths in file transfer logs. Hypervisor-level logging (QEMU/KVM event logs) may record unusual file transfer requests. Check guest system file integrity monitoring tools for unexpected changes in critical system directories. Correlate spice-vdagent process creation and file operations with user and host activity to identify potential compromise.

Why prioritize this

While the CVSS score is MEDIUM (4.4), prioritization should reflect your virtualization trust model. Organizations with strict SPICE host access controls can defer patching to normal maintenance windows. Those running SPICE in less-controlled environments—for example, development teams with broad host access, or multi-tenant scenarios—should prioritize higher. The vulnerability poses real risk only if the attacker is already positioned at the host, making it a secondary or tertiary stage in an attack chain rather than a direct entry point.

Risk score, explained

The CVSS 3.1 score of 4.4 reflects: (AV:L) local attack vector—the attacker must be on or control the SPICE host; (AC:L) low attack complexity—once positioned, exploitation is trivial; (PR:H) high privilege requirement—the attacker must be a privileged SPICE host component or process; (UI:N) no user interaction—the guest user need not take action; (S:U) no scope change—impact is limited to the guest system; (C:N) no confidentiality impact; (I:H) high integrity impact—arbitrary file write; (A:N) no availability impact. The MEDIUM rating appropriately weights the serious integrity risk against the high barrier to initial access.

Frequently asked questions

Can an attacker exploit this without controlling the SPICE host?

No. This vulnerability requires the attacker to be in a privileged position within the SPICE infrastructure—either as an administrator of the host or through a prior compromise. Remote, unauthenticated exploitation is not possible.

What is the difference between patching the SPICE host versus patching the guest?

The vulnerability is in spice-vdagent, which runs on the guest. Patching the guest with an updated spice-vdagent version fixes the flaw. Patching or securing the SPICE host is a separate control that limits the likelihood of attack, but does not remediate the guest-side weakness.

Could this vulnerability be used to escalate privileges on the guest?

Not directly. The spice-vdagent process typically runs with user privileges. However, if an attacker writes files to locations used by privileged processes (e.g., cron.d, systemd units with root execution), they could achieve privilege escalation indirectly.

Does this affect non-virtualized systems?

No. This vulnerability requires the SPICE agent and a SPICE host to be present. Systems not running SPICE or not using the spice-vdagent package are unaffected.

This analysis is based on CVE-2026-57966 as published by security databases and vendor advisories. Actual affected product versions and patch availability should be verified directly with Red Hat and the SPICE project before making patching decisions. Exploitation of this vulnerability is contingent on prior access to or compromise of the SPICE host infrastructure and does not represent a standalone remote code execution risk. Organizations should assess their specific SPICE deployment topology and trust model when prioritizing remediation. Source: NVD (public-domain), retrieved 2026-08-08. Analysis generated by SEC.co (claude-haiku-4-5).