# CVE-2026-126996: SakanaAI text-to-LoRA Checkpoint Loading model_loading.py get_tokenizer path traversal

> A crafted model name can make text-to-LoRA load files outside the intended directory; users loading untrusted models should care.

- **Severity:** Medium 5.3 (`CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P`)
- **Exploited in the wild:** Not known
- **Login needed:** Not needed
- **Attackable:** Over the internet
- **Victim action needed:** A user has to click or open something
- **Harmful if only installed:** No
- **Needs a specific call:** Yes (`get_tokenizer`)
- **Fix available:** No
- **Published:** 12 Nov 2026
- **Updated:** 16 Nov 2026
- **Also known as:** GHSA-25cm-vgcw-2gfw

## What is CVE-2026-126996?

CVE-2026-126996 is a medium-severity path traversal vulnerability in SakanaAI text-to-LoRA, published on 12 Nov 2026. An attacker can exploit it over the internet without logging in, and only if a user opens a file or follows a link.

### Official description

A vulnerability was determined in SakanaAI text-to-LoRA 0.0.1. Affected by this issue is the function get_tokenizer of the file src/hyper_llm_modulator/utils/model_loading.py of the component Checkpoint Loading. Executing a manipulation of the argument base_model_name_or_path can lead to path traversal. The attack may be launched remotely. The exploit has been publicly disclosed and may be utilized. The project was informed of the problem early through an issue report but has not responded yet.

Weakness: CWE-22 (Improper Limitation of a Pathname to a Restricted Directory)

## Which versions are affected by CVE-2026-126996?

An affected version is only exploitable if your software calls get_tokenizer. Many projects that include it are not exposed.


## How do I fix CVE-2026-126996?

No fix has been published for CVE-2026-126996 yet. Limit exposure of the affected software until one is available.

## Is CVE-2026-126996 being exploited?

It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-126996 exists.

## Fix prompt for AI coding agents

```text
Check this codebase for CVE-2026-126996 and fix it if it is affected.

What it is: CVE-2026-126996 is a medium-severity path traversal vulnerability in SakanaAI text-to-LoRA, published on 12 Nov 2026. An attacker can exploit it over the internet without logging in, and only if a user opens a file or follows a link.
Exploitation: It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-126996 exists.
Exposure: An affected version is only exploitable if your software calls get_tokenizer. Many projects that include it are not exposed.

Affected product: SakanaAI text-to-LoRA. No package-level version data is available, so check the vendor advisory linked from the reference page.

Task:
1. Search the dependency manifests and lockfiles (for example package.json, package-lock.json, requirements.txt, poetry.lock, uv.lock, pom.xml, build.gradle, go.mod, Cargo.lock, Gemfile.lock) and any Dockerfiles for the affected packages, including transitive dependencies.
   If an affected version is present, also search the code for calls to get_tokenizer and report whether any exist.
2. If an affected version is in use, upgrade it to the fixed version for its release line with the project's package manager, so the lockfile is regenerated. Prefer the smallest upgrade that fixes it.
3. If there is no fix, or the upgrade would break the build, do not guess: explain the risk and the safest workaround instead.
4. Run the tests and summarise what changed and why.
5. If nothing in the project is affected, say so and change nothing.

Reference: https://insights.sbomify.com/CVE-2026-126996/ (Markdown: https://insights.sbomify.com/CVE-2026-126996/index.md)
```

## Sources

- CVE record: https://www.cve.org/CVERecord?id=CVE-2026-126996
- NVD: https://nvd.nist.gov/vuln/detail/CVE-2026-126996
- OSV: https://osv.dev/vulnerability/CVE-2026-126996

---
From sbomify Insights: https://insights.sbomify.com/CVE-2026-126996/
Summaries are generated and may contain mistakes; the official description above is the record.
