# HuggingFace transformers vulnerable to remote code execution

Advisory record · `GHSA-29pf-2h5f-8g72`

- **Identifiers:** `GHSA-29pf-2h5f-8g72` · `CVE-2026-4372` · `PYSEC-2026-2289`
- **State:** Confirmed
- **Severity:** CVSS 7.8 High
- **Critical:** No
- **Weakness (CWE):** [CWE-502](https://cwe.mitre.org/data/definitions/502.html) · [CWE-1066](https://cwe.mitre.org/data/definitions/1066.html)
- **Tags:** [Insecure deserialization](https://agentprime.io/advisories/tag/deserialization.md) · [CWE-1066](https://agentprime.io/advisories/tag/cwe-1066.md)
- **Published:** 26 May 2026
- **Updated:** 10 Sept 2026
- **Reviewed by a person:** No

Published by the GitHub Advisory Database at 26 May 2026, 13:30 UTC. agentprime has not analysed this yet, so whether it affects any app is not known.

## Affected packages

- `transformers` (PyPI): affected `<5.3.0`; fixed in `5.3.0`

## Machine-written summary

Summary written by Claude Haiku from the source advisory. Not reviewed by a person.

> The transformers library, affected before version 5.3.0, can be made to run attacker-supplied Python code. A malicious config.json that sets an internal attention field to an attacker-controlled Hugging Face Hub repository ID causes the library to download and execute code from that repository when a victim loads the model with the standard from_pretrained() call, even without trust_remote_code enabled. Upgrade to transformers 5.3.0 or later.

Written 5 Nov 2026.

## Not analysed yet

We haven’t published our analysis of this Advisory yet. Once a person has reviewed it, this page will say what’s vulnerable, what has to be true for your app to be affected, and how to check.

## Severity and exploitation

- 7.8 High, CVSS 3.0: `CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H`, from CNA (@huntr_ai)

- **CISA KEV:** Not listed.
- **EPSS:** 0.0070, the 53rd percentile, on 11 Nov 2026: FIRST’s estimate of the chance of exploitation in the next 30 days.

## Timeline

- 31 Oct 2026, 14:34 UTC: First seen by agentprime
- 31 Oct 2026, 14:34 UTC: [Confirmed by OSV](https://osv.dev/vulnerability/GHSA-29pf-2h5f-8g72)

## Sources

- [GitHub Advisory Database: GHSA-29pf-2h5f-8g72](https://github.com/advisories/GHSA-29pf-2h5f-8g72)
- [OSV: GHSA-29pf-2h5f-8g72](https://osv.dev/vulnerability/GHSA-29pf-2h5f-8g72)
- <https://nvd.nist.gov/vuln/detail/CVE-2026-4372>
- <https://github.com/huggingface/transformers/commit/a7f8e7ff37d87d1a1a0c8cf607971c607741452f>
- <https://github.com/huggingface/transformers>
- <https://huntr.com/bounties/1f693a6e-6836-4b8b-a0bd-ca036fba8884>

Facts on this page come from the GitHub Advisory Database (CC-BY 4.0), OSV and the other sources listed, as published. We don’t reproduce full advisory descriptions, which can contain exploit code.

## Check whether this affects you

A Match means an installed version is in the affected range above. Most don’t affect the app; agentprime hasn’t analysed this one yet.

- With the agentprime MCP server (<https://api.agentprime.io/mcp>), ask your agent to check the project: `check_lockfile` sends a lockfile, `check_packages` a list of installed packages, and both return the Matches and the upgrades that clear them. `get_advisory` returns this record. Setup: <https://agentprime.io/mcp.md>
- In a browser: <https://agentprime.io/check/> reads your lockfile in the browser, never uploading it.
- Whether it touches your code, with file and line: the Claude Code plugin, <https://agentprime.io/plugin.md>
- As JSON: <https://api.agentprime.io/v1/advisories/GHSA-29pf-2h5f-8g72>, or `POST https://api.agentprime.io/v1/match` with your packages (<https://agentprime.io/api.md>)

---

The web page: <https://agentprime.io/advisories/GHSA-29pf-2h5f-8g72/>. Everything for agents: <https://agentprime.io/llms.txt>.