CVE-2026-70660leejet stable-diffusion.cpp out-of-bounds read
CVE-2026-70660 is a out-of-bounds read in leejet stable-diffusion.cpp, rated medium (6.9 on CVSS 4.0).
Why this priority: Medium severity, with no exploitation signal.
Published · Updated · NVD status: Received
What it is
Code reads past the end of a buffer, leaking memory contents or crashing.
How it is reached
An attacker can exploit it by someone who can already run code or sign in on the machine, without an account, but someone has to act first, such as opening a file or following a link.
What an attacker gains
A successful attack can read the data it protects and take it offline. In practice, this kind of flaw lets an attacker crash the process, read memory, or in the worst case run code.
Is it being exploited?
CISA's SSVC assessment notes a public proof of concept. EPSS estimates a 0.13% chance of exploitation activity in the next 30 days, higher than 3% of scored CVEs.
What to do
- No fixed version is listed yet. Find every place leejet stable-diffusion.cpp runs, watch the references for a fix, and plan a replacement if it is no longer supported.
- Limit who can run code or sign in on the hosts that run it.
- Warn the people who use it, and filter files and links on the way in.
Description from NVD
stable-diffusion.cpp is a pure C/C++ library for running diffusion model (Stable Diffusion, Flux, Wan, Qwen Image, Z-Image, and more) inference. Prior to master-764-e22272e, the safetensors loader in src/model_io/safetensors_io.cpp trusts each tensor's data_offsets after checking only that nbytes equals end - begin. A crafted .safetensors file can declare a byte range beyond the file's data section, causing src/model_loader.cpp to consume an attacker-chosen length from an unchecked mapped-file offset. Loading the model can read adjacent process memory into tensor buffers or crash the process. This issue is fixed in master-764-e22272e.
What this means for your architecture.
A CVE names a product. Whether it matters depends on where that product sits in your systems: what reaches it, which trust boundary it guards and what it can reach once compromised. Veriom maps your components, their relationships and your trust boundaries from your code, cloud and delivery evidence, so you can see where a component like this sits. The radar itself does not look at your systems.
Components it touches
- Native code: operating systems, browsers and firmware. The flaw sits wherever native code reads untrusted data.
- Network appliances and media or file parsers. The flaw sits wherever native code reads untrusted data.
Trust boundaries in play
- Host boundary. The attacker needs a foothold on the machine first; the flaw turns a small foothold into a bigger one.
- People who open files and links. Mail, chat and browsers are the delivery path.
- Wherever native code reads untrusted data. An attacker who crosses it can crash the process, read memory, or in the worst case run code.
Questions to answer about your own systems
- Where does leejet stable-diffusion.cpp run in your systems, and at which version?
- Which of those hosts also run code or users you do not fully trust?
- What can it reach once compromised: which data stores, secrets and services?
Read more
Patches and commits
Sources
- CVE records and CVSS scores from the NVD CVE API. This product uses the NVD API but is not endorsed or certified by the NVD.
- Exploitation status from CISA’s Known Exploited Vulnerabilities catalog.
- Exploit likelihood from FIRST’s Exploit Prediction Scoring System (EPSS), refreshed daily.
- Explanations are written by Veriom from fixed templates over each record’s data, not by a person reviewing each CVE. Descriptions are quoted from the records as published.