CVE-2026-77932
Unauthenticated RCE via /api/v1/chat/react-agent: admin role grants root code execution.
- CVSS 9.3
- CWE-306
- Authentication, Authorization, and Session Management
- Remote
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.
- CWE
- CWE-306
- CVSS base score
- 9.3
- Published
- 2026-10-20
- OWASP
- A01 Broken Access Control
- Orthogonal defect classification
- Interface
- Code defect classification
- Missing Check
- Category
- Authentication, Authorization, and Session Management
- Subcategory
- Privilege Escalation
- Accessibility scope
- Remote
- Impact
- Arbitrary Code Execution
- Affected component
- DB-GPT
Solution
Upgrade DB‑GPT to version 0.8.3 or later, which removes the insecure mock get_user_from_headers implementation and disables the unsandboxed code_interpreter and shell_interpreter tools for untrusted users.
Vulnerable code sample
from flask import Flask, request, jsonify
app = Flask(__name__)
# Mock function that erroneously grants admin to everyone
def get_user_from_headers(headers):
# VULNERABLE: always returns admin role
return {"username": "anonymous", "role": "admin"}
@app.route("/api/v1/chat/react-agent", methods=["POST"])
def react_agent():
user = get_user_from_headers(request.headers)
if user["role"] != "admin":
return jsonify({"error": "Forbidden"}), 403
# Unsandboxed tool execution
from react_agent import ReActAgent
agent = ReActAgent()
result = agent.run(request.json.get("prompt", ""))
return jsonify({"response": result})Patched code sample
from flask import Flask, request, jsonify
app = Flask(__name__)
# Proper authentication: verify token and lookup role
def get_user_from_headers(headers):
# FIX: validate token and assign correct role
token = headers.get("Authorization", "").replace("Bearer ", "")
if token == "secret-admin-token":
return {"username": "admin", "role": "admin"}
elif token == "secret-user-token":
return {"username": "user", "role": "user"}
return None
@app.route("/api/v1/chat/react-agent", methods=["POST"])
def react_agent():
user = get_user_from_headers(request.headers)
if not user or user["role"] != "admin":
return jsonify({"error": "Forbidden"}), 403
# Unsandboxed tool execution
from react_agent import ReActAgent
agent = ReActAgent()
result = agent.run(request.json.get("prompt", ""))
return jsonify({"response": result})Cite this entry
@misc{vaitp:cve202677932,
title = {{Unauthenticated RCE via /api/v1/chat/react-agent: admin role grants root code execution.}},
author = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
year = {2026},
note = {VAITP Python Vulnerability Dataset, entry CVE-2026-77932},
howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-77932/}}
}
Introducing the "VAITP dataset": a specialized repository of Python vulnerabilities and patches, meticulously compiled for the use of the security research community. As Python's prominence grows, understanding and addressing potential security vulnerabilities become crucial. Crafted by and for the cybersecurity community, this dataset offers a valuable resource for researchers, analysts, and developers to analyze and mitigate the security risks associated with Python. Through the comprehensive exploration of vulnerabilities and corresponding patches, the VAITP dataset fosters a safer and more resilient Python ecosystem, encouraging collaborative advancements in programming security.
The supreme art of war is to subdue the enemy without fighting.
Sun Tzu – “The Art of War”
:: Shaping the future through research and ingenuity ::
