/

CVE-2026-48769

Severity:

CRITICAL RISK

(9.9)

Improper Input Validation (CWE-20) in Incus Image Client

CVE-2026-48769

Deferred

Vulnerability type

Category:

Improper Input Validation

CWE:

CWE-20

Impact:

Arbitrary File Write and Root-Level Command Execution

Severity:

CRITICAL RISK

(9.9)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A critical input-validation vulnerability in the Incus client allows a malicious image server to supply a crafted Incus-Image-Hash header that is processed unsafely. Under affected conditions, this may enable arbitrary file writes and could result in command execution with root privileges on the Incus server.

Affected context

The vulnerability occurs in Incus deployments that use an affected client to retrieve images from an image server controlled or compromised by an attacker. Risk depends on how the client processes the returned Incus-Image-Hash header and whether the resulting file-write behavior reaches a privileged server-side execution path.

Exploit preconditions

  • Attack surface: Incus image retrieval and image-server communication
  • Required access: The attacker must control, compromise, or impersonate an image server trusted by the Incus client
  • Trigger point: Processing of a malicious Incus-Image-Hash response header
  • Why exploitation is possible: Insufficient validation of header-derived data can allow attacker-controlled values to influence file paths or file contents, potentially reaching privileged execution contexts

Why this happens

The underlying bug is improper validation of externally supplied image metadata. The client accepts a response header from an image server and uses its value in file-handling logic without sufficiently constraining its format, destination, or permitted effects.

This maps to CWE-20: Improper Input Validation, where data is not adequately checked against the expected syntax, structure, or security policy before use. When untrusted metadata influences privileged file operations, validation failures may become arbitrary file-write and code-execution risks.

Potential impact

  • An attacker may write or overwrite arbitrary files accessible to the affected process
  • The attacker could execute commands with root privileges on the Incus server
  • The attacker may compromise containers, virtual machines, or host management functions
  • System integrity, confidentiality, and availability could be affected
  • The vulnerability could support persistence or lateral movement within the infrastructure

General mitigation guidance

Mitigation generally focuses on reducing trust in image-server responses and constraining all metadata-driven file operations.

Common defensive patterns include:

  • Applying the vendor-provided security fix or an equivalent corrected implementation
  • Validating image headers against a strict, allowlisted format before use
  • Treating header values as data rather than filesystem paths, commands, or executable content
  • Resolving file destinations within fixed, non-privileged directories and rejecting traversal or special path forms
  • Separating image retrieval from privileged host operations
  • Running image-processing components with the minimum required permissions
  • Restricting trusted image servers through authentication, integrity verification, and network policy
  • Using read-only or isolated staging areas for downloaded image material

How teams detect issues like this

High-level detection can combine code analysis, runtime monitoring, and trust-policy enforcement:

  • Use static analysis to identify response-header values flowing into filesystem APIs, process execution, or privileged operations
  • Review code for missing allowlists, canonical-path checks, destination constraints, and validation before file writes
  • Monitor Incus clients for unexpected image-server endpoints or unusual Incus-Image-Hash values
  • Alert on writes to sensitive host paths or unexpected child processes launched by Incus components
  • Enforce policies requiring authenticated and approved image sources
  • Use automated security review tools to flag untrusted input reaching file and command-handling APIs

Sources

SourceLink
github.comgithub.com/canonical/lxd/security/advisories/GHSA-pjff-c2wc-f6jm
github.comgithub.com/lxc/incus/security/advisories/GHSA-f6m5-xw2g-xc4x

Frequently asked

questions

It may be exploitable through the network when an Incus client retrieves images from an attacker-controlled, compromised, or improperly trusted image server. The exact exposure depends on image-server configuration and deployment trust relationships.

The attacker generally needs the ability to influence the image server or its responses and must have the target client use that server for image operations. Direct host access is not necessarily required.

Use the corrected vendor implementation and ensure that image metadata is strictly validated before it can influence file operations. File destinations should be fixed, constrained, and separated from privileged execution paths.

Treat all remote image metadata as untrusted, authenticate and restrict image sources, apply least privilege, monitor sensitive file writes, and use static or automated review checks for untrusted data flowing into filesystem and process APIs.

Find out what's already
exploitable in your codebase.