CVE-2026-92702

CriticalCVSS 9.1

Published: 18/09/2026(UTC)Translated: 19/09/2026(UTC)NVD NIST

Summary

Cocos AI up to version 0.8.2 has a vulnerability in the aTLS verification path for AMD SEV-SNP where attestation freshness is not enforced when the expected reportData is nil, empty, or omitted. This allows acceptance of unrelated or stale evidence, potentially leading to trust in an unintended attestation context.

Risk Assessment

A relying party using this path can be tricked into trusting an unintended context, potentially leading to confidentiality or integrity compromise of data processed in the TEE.

Recommendation

Upgrade Cocos AI to version 0.9.0 or later. Ensure a non-empty reportData is always provided.

Other vulnerabilities in Cocos AI

CVE-2026-92701Critical
Cocos AI up to version 0.8.2 has a vulnerability in the aTLS verification path for Intel TDX where the expected freshness value is not copied into the quote-body policy before validation. This allows acceptance of evidence with mismatched or reused reportData, potentially leading to session misbinding.

Original NVD description (English source)

Cocos AI is a confidential computing system for running AI workloads inside trusted execution environments. In versions up to and including 0.8.2, the intra-handshake attested TLS (aTLS) AMD SEV-SNP verification path does not enforce attestation freshness when the expected reportData value is nil, empty, or omitted, leaving the SEV-SNP policy ReportData unset so the verifier accepts unrelated or stale Evidence not bound to the current connection. A relying party that uses this path without an expected reportData as a trust or authorization decision can be induced to trust an unintended attestation context; a supplied non-empty reportData is still validated. The issue is fixed in version 0.9.0.