CVE-2026-129092 in attested-tls-poc
Summary
by MITRE • 11/13/2026
CCC Attestation SIG attested-tls-poc through commit 731ca82 contains a time-of-check time-of-use vulnerability because Evidence is appraised only once during the TLS handshake, without post-handshake re-attestation. Attackers with local access to the attesting endpoint can alter workload code, data, configuration, or attached devices like GPUs while peers keep trusting the original Evidence.
If you want to get best quality of vulnerability data, you may have to visit VulDB.
Analysis
by VulDB Data Team • 11/13/2026
The identified vulnerability represents a critical flaw in the implementation of Time-of-Check-Time-of-Use (TOCTOU) logic within the CCC Attestation SIG attested-tls-poc codebase, specifically as it existed up to commit 731ca82. This architectural deficiency stems from the fact that cryptographic evidence is appraised and validated only once during the initial phase of the Transport Layer Security handshake process. By failing to implement post-handshake re-attestation mechanisms, the system creates a persistent window of opportunity for adversaries who have gained local access to the attesting endpoint. In this scenario, the security model relies entirely on the integrity of the environment at the precise moment of connection establishment, ignoring any subsequent modifications that may occur while the secure channel remains active and trusted by remote peers.
From a technical perspective, the core issue lies in the static nature of the evidence validation process. Once the TLS handshake completes successfully based on the initial appraisal of the Evidence, the peer assumes continuous trust in the attesting endpoint's state. However, because there is no mechanism to re-verify this evidence after the connection is established, an attacker with local privileges can manipulate various components of the system without disrupting the ongoing encrypted session. This includes altering workload code, modifying sensitive data, changing configuration files, or tampering with attached hardware devices such as Graphics Processing Units. The remote peer remains unaware that the underlying environment has been compromised because the cryptographic proof presented at handshake no longer reflects the current state of the endpoint.
The operational impact of this vulnerability is severe, particularly in high-security environments where continuous trust verification is required to prevent data exfiltration or unauthorized execution of malicious code. An attacker leveraging local access can effectively hijack the secure communication channel by injecting malicious logic into running workloads or extracting sensitive information from memory and storage while the connection appears legitimate to external observers. This undermines the fundamental purpose of remote attestation, which is to ensure that software runs in a trusted environment throughout its execution lifecycle rather than just at startup. The ability to alter attached devices like GPUs also introduces risks related to hardware-level attacks, where malicious firmware or drivers could intercept or modify data streams without detection by higher-layer security protocols.
This vulnerability aligns with CWE-367, which defines Time-of-check Time-of-use (TOCTOU) race conditions as a class of software weakness resulting from the failure to properly synchronize access to shared resources when multiple processes are involved. Furthermore, it relates to MITRE ATT&CK technique T1059, Command and Scripting Interpreter, if the attacker uses local modifications to execute further malicious actions within the trusted context established by the flawed attestation process. The lack of continuous verification violates best practices for zero-trust architectures, which mandate that trust must be continuously validated rather than assumed based on initial authentication events.
To mitigate this vulnerability, it is essential to implement post-handshake re-attestation protocols where evidence is periodically reappraised during long-lived connections. This ensures that any changes made by local attackers are detected and the connection is terminated before sensitive operations can proceed under false trust assumptions. Additionally, enforcing strict integrity checks on attached devices like GPUs through hardware-enforced isolation mechanisms such as IOMMU or virtualization-based security features can reduce the attack surface for device-level tampering. Organizations should also consider adopting continuous attestation frameworks that align with industry standards for secure enclaves and trusted execution environments to ensure end-to-end trust throughout the entire lifecycle of a workload rather than relying solely on initial handshake validation.