CVE-2025-11490
Published: 08 October 2025
Summary
CVE-2025-11490 is a medium-severity Command Injection (CWE-77) vulnerability in Wonderwhy-Er Desktopcommandermcp. Its CVSS base score is 6.3 (Medium).
Operationally, exploitation aligns with the MITRE ATT&CK technique Command and Scripting Interpreter (T1059); ranked at the 38.4th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog; a public proof-of-concept is referenced.
This vulnerability is AI-related — categorised as AI Agent Protocols and Integrations; in the LLM/Generative AI Risks risk domain.
The strongest mitigations our analysis identified are NIST 800-53 CM-7 (Least Functionality) and SI-10 (Information Input Validation).
Threat & Defense at a Glance
Threat & Defense Details
Mitigating Controls (NIST 800-53 r5)AI
Directly prevents OS command injection by validating inputs to the extractBaseCommand function against malicious absolute paths and shell metacharacters.
Restricts the command manager to least functionality by permitting only simple whitelisted command names without absolute paths, matching vendor-described safe AI workflows.
Reduces impact of successful command injection by enforcing least privilege on the DesktopCommanderMCP process handling remote low-privilege inputs.
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
OS command injection via absolute path bypass in command handler enables remote arbitrary OS command execution (T1210), abuse of command interpreters (T1059), and indirect command execution through the flawed parser (T1202).
NVD Description
A vulnerability has been found in wonderwhy-er DesktopCommanderMCP up to 0.2.13. The affected element is the function extractBaseCommand of the file src/command-manager.ts of the component Absolute Path Handler. Such manipulation leads to os command injection. The attack may be performed…
more
from remote. The exploit has been disclosed to the public and may be used. The vendor explains: "The usual use case is that AI is asked to do something, picks commands itself, and typically uses simple command names without absolute paths. It's curious why a user would ask the model to bypass restrictions this way. (...) This could potentially be a problem, but we are yet to hear reports of this being an issue in actual workflows. We'll leave this issue open for situations where people may report this as a problem for the long term."
Deeper analysisAI
CVE-2025-11490 is an OS command injection vulnerability affecting wonderwhy-er DesktopCommanderMCP versions up to 0.2.13. The issue resides in the extractBaseCommand function within the src/command-manager.ts file of the Absolute Path Handler component. Manipulation of this function enables arbitrary OS command execution, as classified under CWE-77 and CWE-78, with a CVSS v3.1 base score of 6.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L).
The vulnerability can be exploited remotely by attackers who possess low privileges (PR:L), requiring network access with low complexity and no user interaction. Successful exploitation allows limited impacts on confidentiality, integrity, and availability, potentially enabling command injection to execute arbitrary operating system commands on the affected system.
Advisories reference GitHub issue #218 in the DesktopCommanderMCP repository, where the vendor acknowledges the vulnerability but notes no reports of real-world issues in workflows. They describe the typical use case as AI models selecting simple command names without absolute paths and question user intent to bypass restrictions this way, opting to leave the issue open for potential future reports. No patches are mentioned in the provided references.
The exploit has been publicly disclosed and may be usable. The vulnerability carries AI relevance, as the vendor highlights AI-driven command selection in standard operations, with no known real-world exploitation reported to date.
Details
- CWE(s)
Affected Products
AI Security AnalysisAI
- AI Category
- AI Agent Protocols and Integrations
- Risk Domain
- LLM/Generative AI Risks
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- DesktopCommanderMCP is a component designed for AI agents to execute desktop commands, where AI models generate commands that are processed by the command manager (src/command-manager.ts). The vulnerability involves bypassing a command blocklist via absolute paths in AI-generated commands, directly tying it to AI agent integrations and protocols for command execution.