Live › vulnerability › CVE-2026-77932
Go to dashboardDB-GPT's /api/v1/chat/react-agent endpoint treats every caller as an admin because its user check is a mock, so anyone on the network can send prompts that drive unsandboxed Python and bash tools. An attacker gets code execution as root in the webserver container and can reach the bundled database with default credentials.
Summary written from the sources below. Check the originals before acting.
First seen by InfraStack 2026-10-23 21:57 UTC. All times on this page are UTC.
| Product | Ecosystem | Affected | Fixed in |
|---|
| DB-GPT | pypi | <= 0.8.2 | – |
0.8.2 or earlier (pip show db-gpt, or the image tag in your Docker setup)0.8.2 (fixed version not stated in the sources)Each row is the first time that source carried this signal. The analyst’s rows are its own findings.
DB-GPT through 0.8.2 contains an unauthenticated remote code execution vulnerability because the /api/v1/chat/react-agent endpoint relies on a mock get_user_from_headers function that grants every caller the admin role. Unauthenticated attackers can send prompts that make the ReAct agent's unsandboxed code_interpreter and shell_interpreter tools run arbitrary Python or bash as root in the webserver container, including reaching the bundled database with default credentials.
This page as .md · .json · in your agent: get_signal("CVE-2026-77932")
NVD data is provided by NIST and is public domain. Posts quoted belong to their authors.