CVE-2026-4035 Vulnerability Analysis & Exploit Details

CVE-2026-4035 Vulnerability Summary

CVE-2026-4035: A vulnerability in mlflow/mlflow versions prior to 3.11.0 allows for the resolution of environment variables in AI Gateway secrets, which can be exploited to exfiltrate sensitive server-side environment credentials to an attacker-controlled endpoint. This issue arises because the `api_key` field in gateway secrets can accept `$ENV_VAR` references, which are resolved against the MLflow server's environment during runtime. The resolved secrets are then sent in provider authentication headers to the configured upstream `api_base`. This vulnerability can be exploited by low-privileged authenticated users in basic-auth deployments or by unauthenticated users in default deployments without `basic-auth`. The impact includes potential leakage of sensitive credentials such as cloud artifact credentials (`AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`), which could lead to artifact poisoning and cross-boundary code execution in downstream environments. The issue is fixed in version 3.11.0.

CVE-2026-4035 References

External References

• https://github.com/mlflow/mlflow/commit/4a3f2f720cb4f058c9e0c5b883e0acc9ab64a7f3
• https://huntr.com/bounties/f8e591a0-0f19-4910-b82e-16c9956f2233

CWE Common Weakness Enumeration

CAPEC Common Attack Pattern Enumeration and Classification

• Choosing Message Identifier CAPEC-12 This pattern of attack is defined by the selection of messages distributed via multicast or public information channels that are intended for another client by determining the parameter value assigned to that client. This attack allows the adversary to gain access to potentially privileged information, and to possibly perpetrate other attacks through the distribution means by impersonation. If the channel/message being manipulated is an input rather than output mechanism for the system, (such as a command bus), this style of attack could be used to change the adversary's identifier to more a privileged one.
• Exploiting Incorrectly Configured SSL/TLS CAPEC-217 An adversary takes advantage of incorrectly configured SSL/TLS communications that enables access to data intended to be encrypted. The adversary may also use this type of attack to inject commands or other traffic into the encrypted stream to cause compromise of either the client or server.
• WiFi MAC Address Tracking CAPEC-612 In this attack scenario, the attacker passively listens for WiFi messages and logs the associated Media Access Control (MAC) addresses. These addresses are intended to be unique to each wireless device (although they can be configured and changed by software). Once the attacker is able to associate a MAC address with a particular user or set of users (for example, when attending a public event), the attacker can then scan for that MAC address to track that user in the future.
• WiFi SSID Tracking CAPEC-613 In this attack scenario, the attacker passively listens for WiFi management frame messages containing the Service Set Identifier (SSID) for the WiFi network. These messages are frequently transmitted by WiFi access points (e.g., the retransmission device) as well as by clients that are accessing the network (e.g., the handset/mobile device). Once the attacker is able to associate an SSID with a particular user or set of users (for example, when attending a public event), the attacker can then scan for this SSID to track that user in the future.
• Cellular Broadcast Message Request CAPEC-618 In this attack scenario, the attacker uses knowledge of the target’s mobile phone number (i.e., the number associated with the SIM used in the retransmission device) to cause the cellular network to send broadcast messages to alert the mobile device. Since the network knows which cell tower the target’s mobile device is attached to, the broadcast messages are only sent in the Location Area Code (LAC) where the target is currently located. By triggering the cellular broadcast message and then listening for the presence or absence of that message, an attacker could verify that the target is in (or not in) a given location.
• Signal Strength Tracking CAPEC-619 In this attack scenario, the attacker passively monitors the signal strength of the target’s cellular RF signal or WiFi RF signal and uses the strength of the signal (with directional antennas and/or from multiple listening points at once) to identify the source location of the signal. Obtaining the signal of the target can be accomplished through multiple techniques such as through Cellular Broadcast Message Request or through the use of IMSI Tracking or WiFi MAC Address Tracking.
• Analysis of Packet Timing and Sizes CAPEC-621 An attacker may intercept and log encrypted transmissions for the purpose of analyzing metadata such as packet timing and sizes. Although the actual data may be encrypted, this metadata may reveal valuable information to an attacker. Note that this attack is applicable to VOIP data as well as application data, especially for interactive apps that require precise timing and low-latency (e.g. thin-clients).
• Electromagnetic Side-Channel Attack CAPEC-622 In this attack scenario, the attacker passively monitors electromagnetic emanations that are produced by the targeted electronic device as an unintentional side-effect of its processing. From these emanations, the attacker derives information about the data that is being processed (e.g. the attacker can recover cryptographic keys by monitoring emanations associated with cryptographic processing). This style of attack requires proximal access to the device, however attacks have been demonstrated at public conferences that work at distances of up to 10-15 feet. There have not been any significant studies to determine the maximum practical distance for such attacks. Since the attack is passive, it is nearly impossible to detect and the targeted device will continue to operate as normal after a successful attack.
• Compromising Emanations Attack CAPEC-623 Compromising Emanations (CE) are defined as unintentional signals which an attacker may intercept and analyze to disclose the information processed by the targeted equipment. Commercial mobile devices and retransmission devices have displays, buttons, microchips, and radios that emit mechanical emissions in the form of sound or vibrations. Capturing these emissions can help an adversary understand what the device is doing.

Other 5 Recently Published CVEs Vulnerabilities

• CVE-2026-4035 – A vulnerability in mlflow/mlflow versions prior to 3.11.0 allows for the resolution of environment variables in AI Gateway secrets, which can be ex...
• CVE-2025-15654 – Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in Fox-themes Prague allows Reflected XSS. This...
• CVE-2026-5078 – Impact: The morgan logging middleware's :remote-user token extracts the Basic auth username from the Authorization request header and writes it to ...
• CVE-2026-50052 – In Vinyl Cache before 9.0.1 and Varnish Cache before 9.0.3, a deficiency in HTTP/2 request parsing can be exploited to launch a backend request des...
• CVE-2026-50031 – ipmi-oem in FreeIPMI before 1.6.18 has exploitable buffer overflows on response messages. The Intelligent Platform Management Interface (IPMI) spec...