CVE-2026-7482: Ollama heap out-of-bounds read in GGUF tensor parsing leaks server process memory to unauthenticated remote attackers

Is it exploited?

Reported, not confirmed by CISA. VulnCheck KEV lists it as exploited in the wild. CISA has not added it to its Known Exploited Vulnerabilities catalog as of 29 Oct 2026.

VulnCheck KEV lists it since 8 Jun 2026. It cites 2 public reports, the first from 8 Jun 2026.

• posts.inthecyber.com
• hunt.io

How likely is exploitation in the next 30 days?

2.2%

Up from 0.7% on 23 Oct 2026.

EPSS score, percentile 82.8 of all scored records. A probability estimated by FIRST, not a report of exploitation.

How severe?

8.8 High

CVSS v4.0 base score, given by the CNA

What should I do?

Upgrade to ollama 0.17.1 or later. Full textSolution from the CVE record.

CISA has set no required action, because the vulnerability is not in KEV.

Advisories and patches (3)

Which versions are affected?

Affected

• before 0.17.1

Versions by product

Source: CISA KEV 2026.10.29, as of 29 Oct 2026; VulnCheck KEV, as of 30 Oct 2026; FIRST EPSS, as of 30 Oct 2026; CVE List, as of 30 Oct 2026. About the data.

Description

Ollama heap out-of-bounds read in GGUF tensor parsing leaks server process memory to unauthenticated remote attackers

Ollama before 0.17.1 contains a heap out-of-bounds read vulnerability in the GGUF model loader. The /api/create endpoint accepts an attacker-supplied GGUF file in which the declared tensor offset and size exceed the file's actual length; during quantization in fs/ggml/gguf.go and server/quantization.go (WriteTo()), the server reads past the allocated heap buffer. The leaked memory contents may include environment variables, API keys, system prompts, and concurrent users' conversation data, and can be exfiltrated by uploading the resulting model artifact through the /api/push endpoint to an attacker-controlled registry. The /api/create and /api/push endpoints have no authentication in the upstream distribution. Default deployments bind to 127.0.0.1, but the documented OLLAMA_HOST=0.0.0.0 configuration is widely used in practice (large public-internet exposure observed).

Source: CVE List, as of 30 Oct 2026. About the data.

What to do

Advisories and patches

• ollama/ollama PR #14406 — ggml: ensure tensor size is valid (fix) (Patch)
• Fix commit 88d57d0 (Patch)
• ollama v0.17.1 release notes (Release notes)

Solution (from the CVE record)

Upgrade to ollama 0.17.1 or later. The fix in PR #14406 validates that declared tensor offset+size do not exceed the GGUF file size before reading, and adds a length check in the quantizer prior to the unsafe read.

Workaround (from the CVE record)

Until upgrade is possible: (1) ensure Ollama is bound to a trusted interface only (default OLLAMA_HOST=127.0.0.1); (2) front Ollama with a reverse proxy that requires authentication on /api/create and /api/push; (3) restrict outbound network egress from the Ollama host to prevent exfiltration via /api/push to attacker-controlled registries.

This summary is based on the CVE record as of 30 Oct 2026. Check the vendor's advisory before you act: affected versions and fixes can change.

Affected products and versions

Source: CVE List, as of 30 Oct 2026. About the data.

Timeline