Live Exposure Defense: From CVE to Confirmed Exposure in 12 Hours – See more

New CVE Detected

CVE-2026-65315 – Remote Denial of Service via Uncontrolled Memory Allocation in GGUF Parser – Oll…

Be the first to know when new zero-days emerge:

Summary

CVE-2026-65315 is a high-severity uncontrolled memory allocation vulnerability (CWE-789) in Ollama’s GGUF metadata parser, scored CVSS 4.0: 8.7 / CVSS 3.1: 7.5 (HIGH). An unauthenticated remote attacker can upload a crafted GGUF file as small as 32–40 bytes via Ollama’s model ingestion API endpoints to trigger unrecoverable Go runtime out-of-memory fatal errors or makeslice panics that bypass recovery middleware, crashing the entire server process. All Ollama versions through HEAD commit f0078ae are affected.

Technical details

  • Root cause: The GGUF metadata parser in fs/ggml/gguf.go uses attacker-controlled length and count fields — including string lengths, tensor dimension counts, and metadata array counts — as direct memory allocation sizes via make() calls, without first validating them against the remaining file size (CWE-789: Memory Allocation with Excessive Size Value).
  • Three distinct vulnerable allocation paths exist: (1) readGGUFString allocates a []byte of the declared string length unconditionally for every key, string value, and array element; (2) gguf.Decode reads a 32-bit tensor dimension count up to 0xFFFFFFFF and allocates up to ~34 GB via make([]uint64, dims), bypassing maxArraySize restrictions; (3) readGGUFArray / newArray is called with an uncapped maxArraySize of -1 on the model create path, allocating make([]T, int(n)) with no upper bound.
  • Attack vector: Network-accessible, no authentication or user interaction required. An attacker sends a POST /api/blobs/:digest followed by POST /api/create with a crafted GGUF payload. The vulnerability is also reachable via POST /api/pull using a malicious model registry.
  • Impact: Allocation failures produce either a Go runtime fatal error: runtime: out of memory — uncatchable by recover() or gin middleware — or a makeslice: len out of range panic on non-HTTP goroutines. Both outcomes terminate the entire Ollama server process with no automatic recovery, resulting in a complete and persistent denial of service.

Affected software

  • Ollama — all versions through HEAD commit f0078ae (all versions prior to the fix introduced in PR #17062, merged July 6, 2026)

Severity

  • CVSS 4.0: 8.7 (HIGH) — CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N
  • CVSS 3.1: 7.5 (HIGH) — CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Mitigation and recommended actions

  • Upgrade: A fix was merged into the Ollama main branch via PR #17062 ("create: harden GGUF create flows") on July 6, 2026. Users should upgrade to the latest available Ollama release to receive this patch.
  • Network-level mitigation (if immediate upgrade is not possible): Ollama binds to 0.0.0.0 on port 11434 by default with no authentication enabled. Apply firewall rules or network access controls to restrict inbound access to this port to trusted hosts only. This is the primary interim defense against unauthenticated remote exploitation.

IONIX Status

The IONIX research team is tracking ongoing exploitation attempts and recommends immediate patching. Potentially affected assets are outlined in this post.

References

Are you exposed?

Get a free report of your organization’s exposure to this CVE and threat

How IONIX’s External Exposure Management Platform Detects and Validates
Zero-Days to Shrink MTTR

1

Map your entire attack surface (continously)

IONIX uses multi-factor discovery methods, including DNS analysis, certificate mapping, metadata inspection, and more, to automatically map every internet-facing asset across your environment. This includes cloud instances, third-party platforms, shadow IT, and even forgotten infrastructure that traditional tools miss.

2

Monitor for new CVEs

Dozens of threat intel feeds using agentic technology are continuously analyzed to detect the appearance of proof-of-concept code, exploit kits, and indicators of active targeting. IONIX goes further by applying AI to proactively evaluate whether emerging vulnerabilities are likely to be exploited, even before PoCs go public.

3

Identify Potential External Exposures

Not all CVEs matter. IONIX filters vulnerabilities by asking attacker-centric questions: Can it be reached from the internet? Does it require authentication? Is it being exploited in the wild? This dramatically reduces noise and focuses teams on threats that can actually be weaponized.

4

Create Safe, Scalable Exploit Validations

IONIX transforms real-world PoCs into safe, non-intrusive test payloads that can be run in production environments without disruption. These simulations are precisely targeted to the systems that are vulnerable, ensuring rapid validation without unnecessary load.

5

Execute Exploit Validations

By combining context about software stack, versioning, exposure status, and reachability, IONIX ensures that only the right payloads are executed against the right assets, maximizing efficiency and minimizing risk.

6

Drive Fast and Actionable Remediation

Results are routed through integrations with ticketing, SOAR, and SIEM tools. Issues are written in plain language, bundled into remediation clusters, and prioritized based on asset criticality, exploitability, and blast radius. This shortens mean time to remediation (MTTR) and empowers teams to act with confidence.

Are you exposed?

Get a free report of your organization’s exposure to this CVE and threat

Subscribe to Threat Center RSS

Copy/paste the link below into your preferred RSS reader or follow these instructions to subscribe to Slack alerts.

Get Real-Time CVE Alerts to Your Email

Be the first to know when new zero-days emerge