CVE-2026-103764: Untrusted Pointer Dereference in kvcache-ai Mooncake
CVE-2026-103764 is a critical vulnerability in the Mooncake transfer engine before version 0.3.13. It involves an untrusted pointer dereference in the ServerSession::readHeader function, allowing unauthenticated attackers to read and write arbitrary process memory via the TCP transport data port. Exploitation can disclose sensitive KV cache contents, prompts, secrets, or corrupt memory potentially leading to code execution.
AI Analysis
Technical Summary
Mooncake transfer engine versions prior to 0.3.13 contain an untrusted pointer dereference vulnerability in the ServerSession::readHeader function. This flaw allows unauthenticated remote attackers to send crafted SessionHeader structures with arbitrary address and size values using READ or WRITE opcodes. By doing so, attackers can read or write arbitrary memory in the process, potentially disclosing sensitive KV cache data or corrupting memory to achieve code execution.
Potential Impact
The vulnerability enables unauthenticated remote attackers to read and write arbitrary process memory, which can lead to disclosure of sensitive information such as KV cache contents, prompts, and secrets. Additionally, memory corruption could be leveraged to execute arbitrary code within the affected process. The CVSS 4.0 score of 9.3 reflects the critical severity and the high impact on confidentiality, integrity, and availability without requiring privileges or user interaction.
Mitigation Recommendations
A fix is available in Mooncake version 0.3.13. Users should upgrade to version 0.3.13 or later to remediate this vulnerability. No other mitigation guidance is provided or necessary once the update is applied.
CVE-2026-103764: Untrusted Pointer Dereference in kvcache-ai Mooncake
Description
CVE-2026-103764 is a critical vulnerability in the Mooncake transfer engine before version 0.3.13. It involves an untrusted pointer dereference in the ServerSession::readHeader function, allowing unauthenticated attackers to read and write arbitrary process memory via the TCP transport data port. Exploitation can disclose sensitive KV cache contents, prompts, secrets, or corrupt memory potentially leading to code execution.
CVSS v4.0
Score 9.3critical
Affected software
kvcache-ai
Mooncake
pkg:github/kvcache-ai/MooncakeRun on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
Mooncake transfer engine versions prior to 0.3.13 contain an untrusted pointer dereference vulnerability in the ServerSession::readHeader function. This flaw allows unauthenticated remote attackers to send crafted SessionHeader structures with arbitrary address and size values using READ or WRITE opcodes. By doing so, attackers can read or write arbitrary memory in the process, potentially disclosing sensitive KV cache data or corrupting memory to achieve code execution.
Potential Impact
The vulnerability enables unauthenticated remote attackers to read and write arbitrary process memory, which can lead to disclosure of sensitive information such as KV cache contents, prompts, and secrets. Additionally, memory corruption could be leveraged to execute arbitrary code within the affected process. The CVSS 4.0 score of 9.3 reflects the critical severity and the high impact on confidentiality, integrity, and availability without requiring privileges or user interaction.
Mitigation Recommendations
A fix is available in Mooncake version 0.3.13. Users should upgrade to version 0.3.13 or later to remediate this vulnerability. No other mitigation guidance is provided or necessary once the update is applied.
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- VulnCheck
- Date Reserved
- 2026-10-01T10:39:47.845Z
- Cvss Version
- 4.0
- State
- PUBLISHED
Threat ID: 6abeed64a43b0b3b89129d32
Added to database: 10/01/2026, 23:31:48 UTC
Last enriched: 10/01/2026, 23:46:11 UTC
Last updated: 10/02/2026, 00:33:59 UTC
Views: 10
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.