CVE-2026-110459

Executive Summary

CVE-2026-110459 is an improper access control vulnerability (CWE-284) in the UniFi Protect AI Key that allows a remote, unauthenticated network-adjacent actor to obtain information, make unauthorized configuration or data changes, and cause a denial-of-service condition on the affected device. The vulnerability carries a CVSS v3.1 base score of 8.6 (High) with an attack vector of Network, Low attack complexity, no privileges required, and no user interaction, indicating that exploitation requires only network reachability to the target. No patch, public proof-of-concept, or exploit code is currently available, and the CVE has not been added to CISA KEV, meaning exploitation in the wild has not been confirmed by authoritative sources; the primary risk is to internet- or network-exposed UniFi Protect AI Key deployments relying on perimeter controls alone for protection.

Technical Analysis

The vulnerability is classified under CWE-284, Improper Access Control, which is a broad category describing cases where a system fails to correctly enforce authorization policy for a resource, function, or interface. In the case of UniFi Protect AI Key, the NVD description states that a malicious actor with network access could exploit this weakness to (1) obtain information, (2) make unauthorized changes, and (3) execute a Denial of Service attack on the device.

Because the CNA-provided description and title do not disclose the specific interface (for example, whether the flaw resides in a web management console, an internal API, an authenticated versus unauthenticated endpoint, or an inter-service communication channel), the report cannot attribute the flaw to a named function, class, endpoint, or file path. Any such specificity would be speculative and is deliberately omitted per accuracy requirements. What can be stated with confidence, based strictly on the CVSS vector and the CWE mapping, is the following:

• The attack surface is network-reachable (AV:N). The vulnerable component accepts requests or messages over a network interface rather than requiring local or physical access.
• No authentication is required to reach the vulnerable condition (PR:N), meaning the access control check that should gate the affected operation is either missing, misconfigured, or bypassable without valid credentials.
• No user interaction is required (UI:N), so the victim does not need to click a link, open a file, or otherwise participate for exploitation to succeed.
• Attack complexity is Low (AC:L), which suggests the attacker does not need to win a race condition, bypass randomized memory protections, or satisfy complex preconditions; the exploitable condition is reachable through straightforward network interaction.
• The scope is Unchanged (S:U), indicating the exploit affects only the vulnerable component's own security authority rather than crossing into other security scopes.

The CIA subscore profile — Confidentiality Low (C:L), Integrity Low (I:L), Availability High (A:H) — is technically informative. It indicates that while information disclosure and unauthorized modification are possible, they are bounded in impact, whereas the availability impact is severe. This pattern is consistent with an access control weakness that exposes limited data (for example, device status, configuration metadata, or telemetry), permits constrained configuration or state changes, and more significantly allows an attacker to disrupt device operation, potentially rendering the UniFi Protect AI Key, and by extension the Protect surveillance workflow it supports, unavailable.

The Improper Access Control class typically manifests in one of several ways: missing authorization checks on an endpoint that should require authentication, broken object-level authorization where any network-reachable caller can act on resources it does not own, or inconsistency between an exposed management interface and the checks enforced on it. The presence of a High Availability impact alongside Low Confidentiality and Integrity impacts is often observed when the vulnerable operation can be invoked repeatedly or can drive the device into a resource-exhaustion or crash state. Without the vendor advisory, the exact trigger cannot be confirmed.

Exploitation Scenarios

Because no public PoC, Nuclei template, or exploit code exists for CVE-2026-110459, the following attack chain is reconstructed at the vulnerability-class level from the NVD description and CVSS vector. No endpoint names, parameters, or code artifacts are fabricated.

1. Reconnaissance / Target Identification An attacker with network access to the target environment enumerates reachable UniFi Protect AI Key devices. This could be achieved through network scanning of common management ports, passive discovery of the device on a shared network segment, or Shodan-style internet-wide scanning of exposed instances. The CVSS AV:N vector confirms that the device exposes a network interface through which the vulnerable operation is reachable.

2. Triggering the Access Control Flaw The attacker issues crafted network requests to the vulnerable interface. Because the CVSS vector specifies PR:N and AC:L, the malicious actor does not need valid credentials and does not need to overcome complex protections. The improper access control condition means the device processes the request without properly verifying that the caller is authorized to perform the requested operation or access the requested resource.

3. Payload / Execution at the Vulnerability-Class Level The attacker leverages the unauthorized request to perform one or more of the impacts described by the vendor: reading information the caller should not be able to obtain, issuing changes that modify device state or configuration without authorization, and/or driving the device into a denial-of-service condition. No specific payload structure can be described because the description does not disclose the request format or the interface in question.

4. Post-Exploitation Impact Successful exploitation yields a Low Confidentiality impact (limited information disclosure), a Low Integrity impact (bounded unauthorized changes), and a High Availability impact (significant disruption to the device). The C:L/I:L pattern suggests the attacker cannot necessarily seize full administrative control, but the A:H impact means the device — and the surveillance capability it provides — can be rendered inoperable. In a security-camera deployment, loss of the AI Key component can degrade or disable video analytics or related Protect functionality, with downstream operational and physical-security consequences.

It should be emphasized: no exploit-in-the-wild activity, threat actor campaign, or public tooling is currently associated with this CVE. The chain above is an inferred, class-level model rather than a confirmed exploit.

Impact Assessment

The CVSS v3.1 base score is 8.6 (High). Interpreting the subscores:

• Confidentiality: Low — some information exposure is possible, but its scope is limited.
• Integrity: Low — unauthorized changes can be made, but their breadth is constrained.
• Availability: High — the most severe dimension; the attacker can cause a denial-of-service condition affecting the device.

Business impact centers on availability and trust in the physical security posture. Organizations using UniFi Protect AI Key devices as part of their surveillance and analytics infrastructure face the risk of degraded or lost monitoring coverage during exploitation. Because the vector requires no credentials and no user interaction (PR:N/UI:N), any network-reachable instance is a candidate target, making network exposure the primary risk amplifier. The blast radius is bounded to the affected device and the Protect functions it supports; scope is Unchanged (S:U), so no cross-boundary compromise of unrelated components is assumed. The combination of Low Confidentiality, Low Integrity, and High Availability impacts means that while full data compromise or persistent control is not implied, intermittent or sustained loss of a physical-security component is a realistic outcome, which carries operational and compliance significance for environments with surveillance obligations.

Detection Guidance

Sigma Rule

title: Suspicious Access to UniFi Protect AI Key Management Interface
id: 7e5c1f2a-3b9d-4c6e-8a1f-2d4b6c8e0f12
status: experimental
description: Detects anomalous or repeated unauthenticated access patterns against UniFi Protect AI Key management services consistent with CVE-2026-110459 improper access control exploitation attempts.
references:
  - https://nvd.nist.gov/vuln/detail/CVE-2026-110459
author: valtersit.com Threat Intelligence
date: 2026/01/01
logsource:
  category: webserver
  product: unifi
detection:
  selection:
    cs-method:
      - GET
      - POST
      - PUT
      - DELETE
    sc-status:
      - 200
      - 401
      - 403
      - 500
  filter_legit:
    src_ip:
      - 127.0.0.1
  condition: selection and not filter_legit
  timeframe: 5m
fields:
  - src_ip
  - cs-method
  - cs-uri-stem
  - sc-status
falsepositives:
  - Legitimate administrative sessions from trusted management hosts
level: medium
tags:
  - attack.t1190
  - cve.2026.110459

YARA Rule

rule CVE_2026_110459_UniFi_Protect_AI_Key_AccessControl
{
    meta:
        description = "Detection heuristic for indicators associated with improper access control exploitation against UniFi Protect AI Key (CVE-2026-110459). No specific payload strings are publicly documented, so this rule targets generic exploitation artifacts."
        author = "valtersit.com Threat Intelligence"
        date = "2026-01-01"
        reference = "https://nvd.nist.gov/vuln/detail/CVE-2026-110459"
        severity = "high"
        cve = "CVE-2026-110459"

    strings:
        $s1 = "UniFi Protect" ascii nocase
        $s2 = "AI Key" ascii nocase
        $s3 = "/api/" ascii nocase
        $s4 = "Unauthorized" ascii nocase
        $s5 = "Access-Control" ascii nocase

    condition:
        (uint16(0) == 0x5A4D or uint32(0) == 0x464C457F) and
        any of ($s1, $s2) and any of ($s3, $s4, $s5)
}

Suricata IDS Rule

alert http any any -> any any (msg:"CVE-2026-110459 UniFi Protect AI Key Improper Access Control Attempt"; flow:established,to_server; http.method; content:"GET"; http.uri; content:"/api/"; http.header; content:"User-Agent"; classtype:attempted-admin; reference:cve,2026-110459; reference:url,nvd.nist.gov/vuln/detail/CVE-2026-110459; sid:2026110459; rev:1;)

Note: Because the specific vulnerable endpoint is not disclosed in the available intelligence, this rule uses a broad generic URI pattern (/api/) as a placeholder for tuning. Analysts should refine the http.uri content match once the vendor advisory identifies the affected interface.

Elastic Detection

sequence by source.ip with maxspan=1m
  [ network where network.protocol == "http" and http.request.method in ("GET", "POST", "PUT", "DELETE") and url.path : "/api/*" ]
  [ network where http.response.status_code in (200, 401, 403, 500) and url.path : "/api/*" ]

Splunk SPL Query

index=web OR index=proxy
| search (uri_path="/api/*" OR uri_path="/*/protect*")
| eval cve="CVE-2026-110459"
| stats count values(http_method) as methods values(status) as status_codes by src_ip uri_path
| where count > 20
| sort - count
| table cve src_ip uri_path methods status_codes count

Wazuh OSSEC Rule

<group name="unifi_protect,cve_2026_110459,">
  <rule id="1104591" level="10">
    <decoded_as>json</decoded_as>
    <field name="url">/api/</field>
    <field name="method">GET|POST|PUT|DELETE</field>
    <description>CVE-2026-110459: Possible UniFi Protect AI Key improper access control exploitation attempt</description>
    <mitre>
      <id>T1190</id>
    </mitre>
    <group>attack,web,</group>
  </rule>
  <rule id="1104592" level="12">
    <if_sid>1104591</if_sid>
    <field name="status">500|503</field>
    <description>CVE-2026-110459: Potential denial-of-service condition following access control exploitation on UniFi Protect AI Key</description>
    <group>attack,web,dos,</group>
  </rule>
</group>

Mitigation & Patch Guidance

Patch confidence: 0/4 — UNPATCHED. No vendor fix, workaround, or patched version has been documented in any tracked source. The guidance below is precautionary and based on the vulnerability class (CWE-284, network-reachable, no auth required).

1. Immediate Mitigations

   • Remove direct network exposure: restrict the UniFi Protect AI Key management/service interface to trusted management VLANs and internal hosts only. Block inbound access from untrusted networks and the public internet at the perimeter firewall.
   • Enforce network segmentation so that only authorized administrative subnets can reach the device.
   • Monitor for anomalous access patterns against the device using the detection logic provided above; repeated 200/401/403 responses on management endpoints from unexpected sources may indicate probing.
   • If the device is not operationally required, consider temporarily disabling the vulnerable service until a patch is available.

2. Patch / Upgrade Steps

   • No patch currently exists. Monitor the vendor's official security advisory channels (Ubiquiti) and the NVD entry for CVE-2026-110459 for an updated fixed firmware/software version.
   • When a fix is published, apply it promptly and verify the deployed version matches or exceeds the patched release.

3. Long-Term Hardening

   • Adopt a zero-trust approach to management interfaces: do not assume internal network reachability implies trust.
   • Maintain an inventory of internet-exposed IoT and physical-security devices and continuously reduce their external attack surface.
   • Integrate CVE monitoring (NVD, vendor advisories) into patch management workflows so that unpatched, network-reachable components are prioritized.
   • Apply least-privilege and defense-in-depth controls around physical-security infrastructure, given the operational consequences of availability loss.

References

• NVD — CVE-2026-110459: https://nvd.nist.gov/vuln/detail/CVE-2026-110459
• CWE-284: Improper Access Control: https://cwe.mitre.org/data/definitions/284.html
• Ubiquiti Security Advisory portal (monitor for vendor fix): https://www.ui.com/security
• FIRST EPSS: Not yet scored for this CVE (no EPSS value available).
• CISA Known Exploited Vulnerabilities Catalog: Not listed (no confirmed exploitation in the wild).

Note: No threat actor campaigns, exploit-in-the-wild evidence, public proof-of-concept, or Nuclei template are currently associated with CVE-2026-110459 based on the intelligence provided.