/

CVE-2026-90703

Severity:

CRITICAL RISK

(9.1)

OS Command Injection (CWE-78) in D-Link DWR-M921 Router

CVE-2026-90703

Deferred

Vulnerability type

Category:

OS Command Injection

CWE:

CWE-77, CWE-78

Impact:

Remote Code Execution

Severity:

CRITICAL RISK

(9.1)

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 OS command injection vulnerability affects the D-Link DWR-M921 router through the folderpath argument handled by /boafrm/formDiskCreateShare. A remote attacker may be able to execute operating-system commands when the affected function processes attacker-controlled input without adequately separating data from shell commands.

Affected context

The vulnerability occurs in the router's disk-sharing functionality when the folderpath parameter is supplied to the affected form handler. Risk is present where the management interface or related service is remotely reachable and the parameter reaches command execution logic without sufficient validation or safe argument handling. The exact exposure depends on deployment configuration, network access controls, and whether authentication is required.

Exploit preconditions

  • Attack surface: The router's web-based disk-sharing form handler at /boafrm/formDiskCreateShare
  • Required access: Network access to the exposed management or service interface; authentication requirements depend on deployment configuration
  • Trigger point: Processing of the folderpath argument
  • Why exploitation is possible: Attacker-controlled input may be incorporated into an operating-system command without reliable separation between data and executable command syntax

Why this happens

The underlying bug is improper neutralization of special elements in externally controlled input used to construct or invoke an operating-system command. This behavior maps most specifically to CWE-78: Improper Neutralization of Special Elements used in an OS Command, with CWE-77: Improper Neutralization of Special Elements in Output Used by a Downstream Component representing the broader command-injection class. The insecure pattern occurs when a file path or similar value is treated as trusted command text rather than as a strictly validated argument passed through a non-shell API.

Potential impact

  • An attacker may execute arbitrary commands with the privileges of the affected router service.
  • The device could be compromised or rendered unavailable.
  • Network traffic, configuration data, or connected storage may be exposed.
  • The router could be used to access or attack other systems on reachable networks.
  • Persistent tampering or unauthorized configuration changes could occur.

General mitigation guidance

Mitigation generally focuses on removing shell interpretation from file-management operations and reducing trust in request parameters.

Common defensive patterns include:

  • Use direct, non-shell process and filesystem APIs with fixed operation mappings.
  • Treat path values as data, apply strict allowlisting, and constrain them to approved directories.
  • Reject unexpected characters, encodings, path traversal components, and ambiguous path representations.
  • Apply authentication, authorization, and network segmentation to administrative interfaces.
  • Run the affected service with the minimum filesystem and process privileges required.
  • Use secure-by-default management exposure settings and restrict administrative access to trusted networks.
  • Add regression tests covering command metacharacters, malformed paths, and alternate encodings.

How teams detect issues like this

Detection can combine source review, runtime monitoring, and administrative policy checks.

  • Static analysis can flag request parameters concatenated into shell commands or passed to shell-evaluating APIs.
  • Code review should examine handlers that transform web input into filesystem paths, command-line arguments, or scripts.
  • Runtime monitoring may identify unexpected child-process creation, unusual command execution, or outbound connections from the router service.
  • Web and device logs can be reviewed for anomalous requests to the disk-sharing endpoint and unusual folderpath values.
  • Policy enforcement can restrict management interfaces to authorized networks and alert on configuration changes or service behavior outside approved baselines.

Automated security review tools may help identify command-construction patterns, but findings should be validated against the actual execution path.

Sources

SourceLink
gist.github.comgist.github.com/H3rmesk1t/24a09e799cd873abde6a2a806a030568
vuldb.comvuldb.com/cve/CVE-2026-90703
vuldb.comvuldb.com/submit/917695
vuldb.comvuldb.com/vuln/403245
vuldb.comvuldb.com/vuln/403245/cti
dlink.comdlink.com/

Frequently asked

questions

The reported issue may be exploited remotely when the affected form handler is reachable over the network. Practical exposure depends on interface binding, authentication, firewall rules, and other deployment controls.

The available description does not establish a universal authentication requirement. Access to the vulnerable interface and any applicable authorization checks determine whether an unauthenticated attacker can reach the trigger point.

Avoid shell-based command construction for file operations. Use direct APIs, strictly validate and constrain paths, and pass values as separate arguments rather than interpolating them into command strings.

Apply least privilege, restrict administrative interfaces, enforce allowlisted input validation, review all user-controlled values that reach process execution, and include command-injection tests in development and security review workflows.

Find out what's already
exploitable in your codebase.