CVE-2026-113392: path traversal in Apmupload

CISO Take

CVE-2026-113392 is an authenticated path traversal in the Apmupload component (/cgi-bin/mainfunction.cgi) of the DrayTek Vigor2960 router, letting a low-privileged remote user manipulate the 'action' argument to write or clobber files outside the intended directory. This is a network edge device, not an AI/ML package: it has no tracked AI packages, no dependents data, and our AI classification (ai_tool) looks like a false positive, so AI-specific blast radius is minimal. It is not in CISA KEV, has no EPSS score, no Nuclei template and no SSVC decision, though the vendor references a public write-up of the exploit case, so the technique is documented. The device is end-of-support and the vendor says only major issues (e.g. unauthenticated flaws) will be reviewed, so any unpatched unit stays exposed. Upgrade to firmware 1.5.1.7 where possible, restrict management interface exposure, and plan replacement of the EoS hardware, particularly if it fronts networks hosting ML workloads.

Sources:NVDVulDBATLASdraytek.com.tw

What is the risk?

Moderate (CVSS 7.4 high, scope changed). Exploitation requires network access and low-privilege authentication, with low complexity and no user interaction. Public proof-of-concept documentation lowers the skill bar, but no KEV listing, EPSS score, or scanner template exists. Relevance to AI/ML systems is indirect: the risk is compromise of perimeter infrastructure that could expose adjacent AI environments. End-of-support status raises long-term risk because future flaws will likely go unfixed.

How does the attack unfold?

Initial Access

Attacker authenticates to the Vigor2960 web interface with low-privilege credentials reachable over the network.

AML.T0049

Exploitation

Crafted 'action' argument to the Apmupload handler in mainfunction.cgi triggers path traversal outside the intended directory.

Impact

Arbitrary files are written or overwritten on the router, enabling config tampering, traffic interception or pivoting toward adjacent AI/ML infrastructure.

Initial Access

Attacker authenticates to the Vigor2960 web interface with low-privilege credentials reachable over the network.

AML.T0049

Exploitation

Crafted 'action' argument to the Apmupload handler in mainfunction.cgi triggers path traversal outside the intended directory.

Impact

Arbitrary files are written or overwritten on the router, enabling config tampering, traffic interception or pivoting toward adjacent AI/ML infrastructure.

How severe is it?

CVSS 3.1

7.4 / 10

EPSS

0.4%

chance of exploitation in 30 days

Higher than 34% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○CISA SSVC: Public PoC

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AVNetwork

ACLow

PRLow

UINone

SChanged

CLow

ILow

ALow

What should I do?

1 step

1) Upgrade Vigor2960 firmware to 1.5.1.7. 2) Disable or restrict WAN-side access to the web management interface and limit it to trusted admin networks or VPN. 3) Rotate router administrator and low-privilege credentials and remove unneeded accounts. 4) Monitor HTTP logs for requests to /cgi-bin/mainfunction.cgi with action parameters containing '../' sequences and unexpected file changes. 5) Plan migration off the end-of-support Vigor2960 to a supported platform. 6) Segment ML infrastructure from the edge device's trust zone.

What does CISA's SSVC say?

DecisionTrack*

Exploitationpoc

AutomatableNo

Technical Impactpartial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Auth BypassFrameworkAML.T0049- Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

ISO 42001

A.4.4-Tooling resources

NIST AI RMF

MANAGE 2.3-Risk treatment and response to unknown risks

Frequently Asked Questions

What is CVE-2026-113392?

CVE-2026-113392 is an authenticated path traversal in the Apmupload component (/cgi-bin/mainfunction.cgi) of the DrayTek Vigor2960 router, letting a low-privileged remote user manipulate the 'action' argument to write or clobber files outside the intended directory. This is a network edge device, not an AI/ML package: it has no tracked AI packages, no dependents data, and our AI classification (ai_tool) looks like a false positive, so AI-specific blast radius is minimal. It is not in CISA KEV, has no EPSS score, no Nuclei template and no SSVC decision, though the vendor references a public write-up of the exploit case, so the technique is documented. The device is end-of-support and the vendor says only major issues (e.g. unauthenticated flaws) will be reviewed, so any unpatched unit stays exposed. Upgrade to firmware 1.5.1.7 where possible, restrict management interface exposure, and plan replacement of the EoS hardware, particularly if it fronts networks hosting ML workloads.

Is CVE-2026-113392 actively exploited?

No confirmed active exploitation of CVE-2026-113392 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-113392?

1) Upgrade Vigor2960 firmware to 1.5.1.7. 2) Disable or restrict WAN-side access to the web management interface and limit it to trusted admin networks or VPN. 3) Rotate router administrator and low-privilege credentials and remove unneeded accounts. 4) Monitor HTTP logs for requests to /cgi-bin/mainfunction.cgi with action parameters containing '../' sequences and unexpected file changes. 5) Plan migration off the end-of-support Vigor2960 to a supported platform. 6) Segment ML infrastructure from the edge device's trust zone.

What systems are affected by CVE-2026-113392?

This vulnerability affects the following AI/ML architecture patterns: Network perimeter / edge routers, Remote access infrastructure supporting ML environments.

What is the CVSS score for CVE-2026-113392?

CVE-2026-113392 has a CVSS v3.1 base score of 7.4 (HIGH). The EPSS exploitation probability is 0.40%.

What is the AI security impact?

Affected AI Architectures

Network perimeter / edge routersRemote access infrastructure supporting ML environments

MITRE ATLAS Techniques

AML.T0049Exploit Public-Facing Application

Compliance Controls Affected

ISO 42001: A.4.4

NIST AI RMF: MANAGE 2.3

What are the technical details?

Original Advisory

A vulnerability was found in DrayTek Vigor2960 1.5.1.6 Revision 8218. This vulnerability affects unknown code of the file /cgi-bin/mainfunction.cgi of the component Apmupload. Performing a manipulation of the argument action results in path traversal. The attack is possible to be carried out remotely. The exploit has been made public and could be used. Upgrading to version 1.5.1.7 is able to resolve this issue. Upgrading the affected component is advised. The vendor explains: "2960 1.5.1.7 has corrected many security issues. And 2960 is EoS with limited support, only major issues would be reviewed(e.g. unauthenticated vuln.)". This vulnerability only affects products that are no longer supported by the maintainer.

Exploitation Scenario

An attacker who has obtained low-privilege credentials for the router (via credential reuse, weak passwords, or phishing) sends a crafted request to /cgi-bin/mainfunction.cgi with a manipulated 'action' parameter in the Apmupload handler containing traversal sequences. This overwrites or plants files outside the upload directory, potentially altering configuration or scripts on the device. From the foothold, the attacker could intercept or redirect traffic to an organization's internal ML services or use the router as a pivot toward model-serving or data-pipeline hosts.

Weaknesses (CWE)

CWE-22Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')Primary CWE-22Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')

CWE-22 — Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'): The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

[Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
[Architecture and Design] For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.

Source: MITRE CWE corpus.