---
id: "CVE-2026-123019"
kind: "vulnerability"
aliases: ["CVE-2026-123019"]
title: "DB-GPT: HuggingFace adapter lets remote input load untrusted code"
severity: "high"
cvss: 7.5
epss: null
exploited: false
status: "open"
category: "ai-tooling"
products: [{"name": "DB-GPT", "fixed": [], "affected": "<= 0.8.2", "ecosystem": "pip"}]
ecosystems: []
tags: ["ai-apps", "ai-tooling", "db-gpt", "huggingface", "python", "self-hosted"]
not_affected_if: ["newer-version", "no-huggingface-adapter"]
published_at: "2026-11-05T12:17:50+00:00"
first_seen_at: "2026-11-05T12:38:47+00:00"
updated_at: "2026-11-05T12:41:17+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-123019", "first_seen_at": "2026-11-05T12:38:47+00:00"}, {"source": "analyst", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-123019", "first_seen_at": "2026-11-05T12:41:17+00:00", "cites": ["https://nvd.nist.gov/vuln/detail/CVE-2026-123019", "https://github.com/eosphoros-ai/DB-GPT/releases"]}]
exploit_sources: []
---

# DB-GPT: HuggingFace adapter lets remote input load untrusted code

DB-GPT, an AI application framework, has a weakness in its HuggingFace adapter where manipulating the trust_remote_code argument can cause untrusted code to be loaded. A remote attacker could use a public exploit to run functionality from an untrusted source on the affected server.

_Summary, check and fix written by a model from the sources; labelled as such._

## Affected

- pip:DB-GPT affected <= 0.8.2 fixed no fixed version listed

## How to check

Check whether DB-GPT is 0.8.2 or earlier (0.8.2 is the latest release, so every release is in range). Look in hf_adapter.py and your model config for models loaded with trust_remote_code enabled, and whether the DB-GPT API is reachable by untrusted users.

## Fix

No fixed version known: the latest release is v0.8.2, and the NVD entry says the project has not responded to the report. Until a patch ships, do not enable trust_remote_code for models, only load models you trust, and keep the DB-GPT API off untrusted networks.

## Timeline

- 2026-11-05T12:38:47+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-123019
- 2026-11-05T12:41:17+00:00 analyst https://nvd.nist.gov/vuln/detail/CVE-2026-123019 — Only NVD/VulDB-derived text found; no maintainer advisory or issue for this CVE located. Latest DB-GPT release is v0.8.2 (26 Aug), the affected ceiling, so no fixed version exists in releases. Nothing verified; PoC link not found.

## Documents

- [describes] NVD CVE-2026-123019 https://nvd.nist.gov/vuln/detail/CVE-2026-123019
- [mentions] DB-GPT releases (latest v0.8.2) https://github.com/eosphoros-ai/DB-GPT/releases

## From the sources

A weakness has been identified in eosphoros-ai DB-GPT up to 0.8.2. Impacted is the function exec_module of the file hf_adapter.py of the component HuggingFace Adapter. Executing a manipulation of the argument trust_remote_code can lead to inclusion of functionality from untrusted control sphere. The attack can be launched remotely. The exploit has been made available to the public and could be used for attacks. The project was informed of the problem early through an issue report but has not responded yet.

---

_Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai_