Horizon Alert
Summary of the vulnerability and why it matters
An unauthenticated attacker can bypass password checks in FastGPT by sending a specially crafted request to the login endpoint. This allows them to log in as any user, including the administrator, potentially compromising the entire system.
- Unauthorized administrator access.
- Affects AI agent building platforms.
- Reachable from the internet.
Attack Path
How an attacker could exploit the issue
An unauthenticated attacker can bypass authentication on FastGPT by exploiting a NoSQL injection flaw in the password login endpoint. By sending a specially crafted MongoDB query operator as the password, an attacker can log in as any user, including the administrator, gaining full control of the platform.
- Attacker targets login endpoint.
- Bypasses password validation.
- Gains admin access.
Live Threat
Current exploitation, exposure, and threat context
This vulnerability allows unauthenticated attackers to bypass password checks and gain administrative access to FastGPT, a platform for building AI agents. Attackers are likely to exploit this for unauthorized control and data access.
- NoSQL injection in login endpoint
- Unauthenticated remote code execution potential
- Internet-facing web application component
Priority actions
Operational Fix
Recommended remediation, mitigation, and detection steps
Prioritize patching FastGPT to version 4.14.9.5 immediately to address the critical NoSQL injection vulnerability. If patching is delayed, implement strict network access controls and monitor for suspicious login attempts and unusual database activity.
- Update FastGPT to 4.14.9.5.
- Block network access if unpatched.
- Monitor for authentication anomalies.
