CVE-2023-37275: System logs spoofable in Auto-GPT via ANSI control sequences
Auto-GPT is an experimental open-source application showcasing the capabilities of the GPT-4 language model. The Auto-GPT command line UI makes heavy use of color-coded print statements to signify different types of system messages to the user, including messages that are crucial for the user to review and control which commands should be executed. Before v0.4.3, it was possible for a malicious external resource (such as a website browsed by Auto-GPT) to cause misleading messages to be printed to the console by getting the LLM to regurgitate JSON encoded ANSI escape sequences (\u001b[). These escape sequences were JSON decoded and printed to the console as part of the model's "thinking process". The issue has been patched in release version 0.4.3.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
Auto-GPTto a version that resolves this vulnerability.Fixed in 0.4.3
Event History
Frequently Asked Questions
What is the severity of CVE-2023-37275?
CVE-2023-37275 has been classified as a high severity vulnerability due to its potential impact on the functionality of the Auto-GPT application.
How do I fix CVE-2023-37275?
To fix CVE-2023-37275, update Auto-GPT to version 0.4.4 or later.
What type of vulnerabilities does CVE-2023-37275 address in Auto-GPT?
CVE-2023-37275 addresses issues related to improper handling of system messages within the Auto-GPT command line UI.
Which versions of Auto-GPT are affected by CVE-2023-37275?
CVE-2023-37275 affects all versions of Auto-GPT up to and including version 0.4.3.
Is CVE-2023-37275 considered a code execution risk?
No, CVE-2023-37275 specifically relates to messaging issues and does not pose a direct code execution risk.