dbcve.org
Llama.cppApplication · Ggml

CVE-2026-21869

CRITICAL · 9.8 CVSS v3.1 Published 2026-01-08
Mitigation only
No fix yet — a mitigation exists. There is no fixed release. A documented workaround reduces exposure in the meantime.
See remediation →
100/100
Remediation priority · Urgent
Public exploit Remotely reachable No privileges Zero-click

Official description Straight from the sourceThe vendor's or NVD's own wording, published unedited. Authoritative, but often terse — it says what broke, rarely what to do.

NVD · unedited
llama.cpp is an inference of several LLM models in C/C++. In commits 55d4206c8 and prior, the n_discard parameter is parsed directly from JSON input in the llama.cpp server's completion endpoints without validation to ensure it's non-negative. When a negative value is supplied and the context fills up, llama_memory_seq_rm/add receives a reversed range and negative offset, causing out-of-bounds memory writes in the token evaluation loop. This deterministic memory corruption can crash the process or enable remote code execution (RCE). There is no fix at the time of publication.

Technical summary Written by usOur analysis, written from the advisory, the CVSS vector and the affected-version data. It adds context the advisory leaves out, and never invents facts that are not in the source.

dbcve analysis · high confidence

llama.cpp server's completion endpoints parse the n_discard parameter directly from JSON without validating it's non-negative. When a negative value is provided and the context fills up, llama_memory_seq_rm/add receives a reversed range and negative offset, causing out-of-bounds memory writes in the token evaluation loop, leading to deterministic memory corruption that can crash the process or enable RCE.

MitigationImplement bounds validation on the n_discard parameter to ensure it is non-negative before use in memory operations, or disable the affected completion endpoints until a patch is available.

Verify against the referenced sources before acting — the references below are authoritative for this CVE, this summary is not.

Affected products & versions What the vendor confirmedThe version ranges the vendor confirmed as vulnerable. If your version sits inside a range here, treat yourself as exposed until you have upgraded.

NVD · CPE data
Llama.cppApplication
Affected:all versions

CVSS breakdown How the score is builtThe industry scoring standard. It rates how the flaw is reached, what it takes to exploit, and what an attacker gains — the score is derived from those, not the other way round.

From the vector
Attack vector
Network
Complexity
Low
Privileges
None
User interaction
None
Scope
Unchanged
Confidentiality
High
Integrity
High
Availability
High

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H

Am I affected? How to checkSteps we derive from the advisory and the affected-version data, so you can decide whether this CVE reaches your setup. They are a guide, not a scan — your own configuration is the authority.

dbcve checks

Work through these to decide whether this CVE applies to you.

  1. Identify if llama.cpp server is running
    Search for llama.cpp server processes using commands like `ps aux | grep llama`, `pgrep -a llama`, or check for processes listening on typical llama.cpp ports (8080, 8081, 8088).
    Affected if The llama.cpp server process is running and exposed.
  2. Confirm the installed version of llama.cpp
    Check the installed package version or binary version by running `llama-server --version`, checking the binary metadata, or reviewing the project's git commit hash if self-compiled.
    Affected if All versions of Ggml Llama.cpp are affected, so any version confirms vulnerability presence.
  3. Verify completion endpoints are accessible
    Check server configuration and exposed HTTP endpoints. Typical completion endpoints include /completion, /v1/completions, or similar paths. Review any reverse proxy or firewall rules that may expose these endpoints.
    Affected if The completion endpoints are network-accessible to untrusted users.
  4. Inspect server startup parameters
    Review the command line or configuration used to start the llama.cpp server, looking for flags that enable the server mode (e.g., --server, -s flag) and port configuration.
    Affected if The server is started with HTTP endpoint exposure enabled.

A defender is affected if they are running any version of llama.cpp server with completion endpoints accessible to users or systems that can send HTTP requests with a crafted n_discard parameter.

Generated from the published advisory. Verify against your own configuration.

Check your environment

Paste your version and any relevant configuration and it will be compared against the affected criteria above. Do not include secrets or credentials.

AI-assisted, checked against the advisory. Informational, not a guarantee.

Remediation Closing itWhat it takes to close this. Where a vendor fix exists we point at it; where none exists we say so plainly, and can build one. Effort estimates are scoped from the advisory, not from your codebase.

dbcve · scoped
Mitigation available No clean upgrade yet — mitigate in the meantime
Mitigation

Implement bounds validation on the n_discard parameter to ensure it is non-negative before use in memory operations, or disable the affected completion endpoints until a patch is available.

Recommended fix High confidence
  1. Monitor the official llama.cpp GitHub repository (github.com/ggerganov/llama.cpp) for security releases addressing the n_discard parameter validation vulnerability.
  2. If possible, implement network-level access controls to limit who can send requests to the llama.cpp server endpoint.
  3. Consider deploying the llama.cpp server behind an API gateway that can perform input validation on the n_discard parameter before forwarding requests.
  4. Review application logs for any unexpected behavior that could indicate exploitation attempts.

Generated from the published advisory — verify against the referenced sources before acting.

Fix this in Llama.cpp Scoped from the published advisory
  • Consultation4.0 h
  • Implementation2.0 h
  • Testing3.0 h
  • Review / QA2.0 h
11.0 hours of engineering $1,970
Get help mitigating

An estimate, not a bill — we confirm scope with you before any work starts. Need it this week? Rush from $3,152.

Scan for this in your stack

Free · runs locally
dbcve dependency scanner

Check whether your project pulls in CVE-2026-21869 — or any other known-vulnerable package — straight from your lock files. Free and open source; it runs locally and uploads nothing.

Get it on GitHub Scan in the browser

References Go to the primary sourcePrimary sources — vendor advisories, patches and trackers. Where our summary and a reference disagree, the reference wins.

Primary sources

Practitioner notes

Contributed

Peer-ranked notes from engineers who’ve handled CVE-2026-21869 in production — separate from our analysis above.

No notes yet

Be the first to add a field note for this CVE — a mitigation you’ve verified, a version caveat, or a link to a working fix. Sign in above to contribute.

What this is

A place for practitioners to share what actually worked: a mitigation you’ve tested, a configuration change, a version- or environment-specific caveat, or a link to a verified patch. The most useful notes rise to the top as peers upvote them, so the signal stays high.

What belongs here
  • Verified mitigations, workarounds, and config changes
  • Version or environment caveats, and links to real fixes
  • No weaponised exploit code, or anything meant to cause harm
  • No spam, self-promotion, credentials, or personal data