PatchSiren

PatchSiren cyber security CVE debrief

CVE-2026-116467 ggml-org CVE debrief

A security flaw has been discovered in ggml-org llama.cpp up to 0.7.0. Affected by this vulnerability is an unknown functionality of the file src/llama-sampler.cpp of the component Mirostat Sampler. The manipulation results in reachable assertion. It is possible to launch the attack remotely. The project was informed of the problem early through a pull request but has not reacted yet.

Vendor
ggml-org
Product
llama.cpp
CVSS
MEDIUM 6.9
CISA KEV
Not listed in stored evidence
Original CVE published
2026-10-26
Original CVE updated
2026-10-27
Advisory published
2026-10-26
Advisory updated
2026-10-27

Who should care

Defenders responsible for ggml-org llama.cpp deployments should assess exposure and verify the affected version. They should also review compensating controls for exposed systems, monitor for potential remote attacks, and track exceptions and retest remediated assets. Security teams and vulnerability management teams should prioritize verifying the affected version and assessing exposure.

Why it matters

Defenders should prioritize verifying the affected version and assessing exposure, as the vulnerability allows for remote attacks and has not been fully addressed by the vendor.

  • Verify affected versions and assess exposure to remote attacks
  • Monitor for potential vulnerabilities in the Mirostat Sampler component
  • Review and update ggml-org llama.cpp deployments to prevent exploitation

Technical summary

The vulnerability is located in the src/llama-sampler.cpp file of the Mirostat Sampler component in ggml-org llama.cpp up to 0.7.0. The manipulation leads to a reachable assertion, allowing for remote attacks. Defenders should prioritize verifying the affected version and assessing exposure, as the vulnerability allows for remote attacks and has not been fully addressed by the vendor. Affected product deployments should be verified, and owners assigned for follow-up. The project was informed of the problem early through a pull request but has not reacted yet.

Defensive priority

Defenders should prioritize verifying the affected version and assessing exposure, as the vulnerability allows for remote attacks.

Recommended defensive actions

  • Verify the version of ggml-org llama.cpp and assess exposure
  • Review the Mirostat Sampler component for potential vulnerabilities
  • Monitor for potential remote attacks
  • Review compensating controls for exposed systems while remediation is scheduled and verified
  • Check relevant monitoring, detection, and logs for exposed assets that need extra review
  • Track exceptions, retest remediated assets, and close the item only after evidence is documented
  • Confirm whether affected product deployments exist in managed environments and assign an owner for follow-up

Evidence notes

The CVE record and NVD entry provide details on the vulnerability in ggml-org llama.cpp up to 0.7.0, specifically in the src/llama-sampler.cpp file of the Mirostat Sampler component. However, the vendor's response and affected versions require further verification. Defenders should verify the affected version, assess exposure, and monitor for potential remote attacks. The manipulation leads to a reachable assertion, allowing for remote attacks. Evidence limits suggest verifying ggml-org llama.cpp deployments and reviewing the Mirostat

Sources and references

Verified primary and authoritative sources

  • CVE-2026-116467 CVE Program record

    Publisher, destination, and source semantics verified

    URL: https://www.cve.org/CVERecord?id=CVE-2026-116467

    CVE Program - Official CVE Program record with source-provided CVE metadata.

  • CVE-2026-116467 NVD vulnerability detail

    Publisher, destination, and source semantics verified

    URL: https://nvd.nist.gov/vuln/detail/CVE-2026-116467

    NIST National Vulnerability Database - Official NIST NVD detail page and source-specific vulnerability assessment.

Supplemental references

Methodology and review provenance

AI-assisted synthesis based on stored public vulnerability evidence. System validation, approval state, and publication status do not by themselves establish human review of this revision. PatchSiren helps prioritize defensive review and does not prove exposure or remediation on any system.