If your organization runs BISHENG — particularly in AI or machine learning infrastructure, where this platform is increasingly common — you need to treat this as an emergency. CVE-2026-130840 is a critical authentication bypass vulnerability with a CVSS 9.8 score, remotely exploitable over the network. It allows unauthenticated attackers to impersonate administrators, granting full access to user accounts, workspaces, knowledge bases, and workflows. This is not a theoretical risk; the flaw is trivial to exploit once an attacker understands the mechanism. Your priority is to identify every BISHENG instance in your environment and patch immediately.
Technical Analysis
Affected Product and Versions:
- BISHENG (all versions before 3.0.0)
Vulnerability Details:
- CVE ID: CVE-2026-130840
- CVSS 3.1 Score: 9.8 (Critical)
- Attack Vector: Network
- Attack Complexity: Low
- Privileges Required: None
- User Interaction: None
How the Vulnerability Works:
BISHENG uses a JWT-based authentication mechanism via the AuthJwt library. The root cause is a hard-coded default jwt_secret embedded in the codebase. This secret is used to sign JSON Web Tokens (JWTs) that grant access to the application. Because the secret is known and static, any attacker can:
- Forge an HS256 Token Offline: Using the hard-coded
jwt_secret, an attacker generates a valid JWT containing auser_idclaim set to1(the default administrator account) and aroleclaim set toadmin. - Send the Token: The forged token is transmitted as an HTTP cookie named
access_token_cookiein requests to the BISHENG application. - Gain Administrative Access: The server validates the token using the hard-coded secret and processes the request with administrative privileges. This allows the attacker to read, modify, or delete any data within the platform, including user credentials, AI model configurations, and proprietary datasets.
Exploitation Status: As of this writing, the vulnerability is publicly documented in the National Vulnerability Database (NVD) with a clear attack path. While there is no confirmed in-the-wild exploitation report yet, the simplicity of the exploit (offline token forgery) means it is highly likely to be weaponized quickly. Security teams should assume active scanning and exploitation attempts are imminent.
Detection & Response
This is a technical threat requiring immediate detection engineering. Below are targeted rules and queries to identify exploitation attempts and post-exploitation activity.
SIGMA Rules
---
title: CVE-2026-130840 BISHENG Admin JWT Forgery Attempt
id: 550d9a23-0f72-4e3a-a8c5-1d2f3b4c5d6e
status: experimental
description: Detects HTTP requests to BISHENG instances containing forged JWT tokens in the access_token_cookie header, indicating potential exploitation of CVE-2026-130840.
references:
- https://nvd.nist.gov/vuln/detail/CVE-2026-130840
author: Security Arsenal
date: 2026/04/06
tags:
- attack.initial_access
- attack.t1078.004
logsource:
category: webserver
product: apache
service: access
detection:
selection:
c-uri|contains:
- '/api/v1/'
- '/auth/'
c-useragent|contains:
- 'python-requests'
- 'curl'
- 'PostmanRuntime'
cs-cookie|contains:
- 'access_token_cookie'
condition: selection
falsepositives:
- Legitimate API clients using cookie-based authentication
level: high
---
title: CVE-2026-130840 BISHENG Admin Role Impersonation
id: 7a3f1c82-9e4b-4d67-bc12-3e5a8f901234
status: experimental
description: Detects HTTP requests to BISHENG with JWT tokens containing admin role claims, which may indicate successful exploitation of CVE-2026-130840.
references:
- https://nvd.nist.gov/vuln/detail/CVE-2026-130840
author: Security Arsenal
date: 2026/04/06
tags:
- attack.privilege_escalation
- attack.t1078
logsource:
category: application
product: bisheng
detection:
selection:
event.type: 'authentication'
event.outcome: 'success'
user.roles|contains: 'admin'
source.ip: '*'
condition: selection
falsepositives:
- Legitimate administrator logins
level: critical
KQL (Microsoft Sentinel / Defender)
// Hunt for potential CVE-2026-130840 exploitation attempts in BISHENG
CommonSecurityLog
| where TimeGenerated > ago(24h)
| where DeviceProduct contains "BISHENG" or DeviceProduct contains "Bisheng"
| where RequestURL contains "/api/v1/" or RequestURL contains "/auth/"
| where RequestCookies contains "access_token_cookie"
| where RequestMethod == "POST" or RequestMethod == "GET"
| summarize by TimeGenerated, SourceIP, DestinationIP, RequestURL, RequestCookies, RequestUserAgent
| order by TimeGenerated desc
// Alternative: Hunt for successful admin logins from external IPs
Syslog
| where TimeGenerated > ago(24h)
| where SyslogMessage contains "login successful" and SyslogMessage contains "admin"
| where SyslogMessage contains "BISHENG" or SyslogMessage contains "bisheng"
| extend SourceIP = extract("src=(\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3})", 1, SyslogMessage)
| where SourceIP !startswith "10." and SourceIP !startswith "192.168." and SourceIP !startswith "172.16."
| summarize by TimeGenerated, SourceIP, SyslogMessage
| order by TimeGenerated desc
Velociraptor VQL
-- Hunt for suspicious processes or network connections on BISHENG servers
SELECT Pid, Name, CommandLine, Exe, Username, CreateTime
FROM pslist()
WHERE CommandLine =~ 'bisheng|authjwt'
OR Exe =~ 'bisheng'
-- Hunt for network connections to BISHENG default ports
SELECT Pid, Name, LocalAddr, LocalPort, RemoteAddr, RemotePort, State
FROM netstat()
WHERE LocalPort in (8080, 8443, 3000)
OR RemotePort in (8080, 8443, 3000)
AND State == 'ESTABLISHED'
Remediation Script (Bash)
#!/bin/bash
# CVE-2026-130840 BISHENG Remediation Script
# This script identifies BISHENG installations, checks version, and applies hardening.
set -euo pipefail
echo "=== CVE-2026-130840 BISHENG Remediation ==="
# Step 1: Find BISHENG installations
echo "[*] Searching for BISHENG installations..."
BISHENG_PATHS=$(find / -type f -name "bisheng*" 2>/dev/null | head -20)
if [ -z "$BISHENG_PATHS" ]; then
echo "[!] No BISHENG installations found."
else
echo "[+] Found potential BISHENG files:"
echo "$BISHENG_PATHS"
fi
# Step 2: Check if BISHENG service is running
echo "[*] Checking for running BISHENG services..."
if systemctl list-units --type=service | grep -i bisheng; then
echo "[!] BISHENG service detected. Stopping for patching..."
sudo systemctl stop bisheng || true
sudo systemctl disable bisheng || true
fi
# Step 3: Backup configuration (if exists)
if [ -d "/opt/bisheng" ]; then
echo "[*] Backing up BISHENG configuration..."
sudo cp -r /opt/bisheng /opt/bisheng_backup_$(date +%Y%m%d_%H%M%S)
echo "[+] Backup created at /opt/bisheng_backup_$(date +%Y%m%d_%H%M%S)"
fi
# Step 4: Check for hard-coded jwt_secret in configuration files
echo "[*] Scanning for hard-coded jwt_secret in BISHENG configs..."
if grep -r "jwt_secret" /opt/bisheng/ 2>/dev/null; then
echo "[!] WARNING: Hard-coded jwt_secret found. Immediate patching required."
else
echo "[+] No hard-coded jwt_secret found in /opt/bisheng/"
fi
# Step 5: Update BISHENG to version 3.0.0 or later
echo "[*] Please update BISHENG to version 3.0.0 or later."
echo " Official vendor advisory: https://github.com/bisheng-ai/bisheng/security/advisories"
echo " Download latest release: https://github.com/bisheng-ai/bisheng/releases"
# Step 6: Restart service after manual update
echo "[*] After updating, restart the service with:"
echo " sudo systemctl start bisheng"
echo " sudo systemctl enable bisheng"
echo "=== Remediation script complete. Manual update required. ==="
Remediation
Immediate Actions:
- Patch Immediately: Upgrade BISHENG to version 3.0.0 or later. The vendor has released a fix that removes the hard-coded
jwt_secretand enforces secure secret generation. Download the latest release from the official repository: https://github.com/bisheng-ai/bisheng/releases - Rotate All Secrets: After patching, generate a new, cryptographically strong
jwt_secretfor each BISHENG instance. Do not reuse secrets across environments. - Audit Access Logs: Review all access logs for the past 30 days for suspicious admin logins, especially from external IP addresses. Look for patterns matching the detection rules above.
- Network Segmentation: Isolate BISHENG instances behind a VPN or internal firewall. Do not expose them directly to the internet unless absolutely necessary.
- Multi-Factor Authentication (MFA): If BISHENG supports MFA, enable it immediately for all administrative accounts. This adds a critical layer of defense even if JWTs are compromised.
Vendor Advisory:
- BISHENG Security Advisory (replace with actual URL when available)
- NVD Entry for CVE-2026-130840
Workarounds (if patching is not immediately possible):
- Disable JWT Authentication: Temporarily switch to a different authentication mechanism (e.g., OAuth2, SAML) if supported.
- WAF Rules: Deploy a Web Application Firewall (WAF) rule to block requests containing
access_token_cookiewith suspicious JWT patterns. - Rate Limiting: Implement strict rate limiting on authentication endpoints to slow down brute-force or token-forgery attempts.
Category
vulnerability-management
Tags
cve-2026-130840, bisheng, auth-bypass, jwt, hardcoded-secret, vulnerability-management
Related Resources
Security Arsenal Penetration Testing Services AlertMonitor Platform Book a SOC Assessment vulnerability-management Intel Hub
Is your security operations ready?
Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.