LowVulnerabilityLLM-specific
CVE-2023-37275: Auto-GPT console spoofing through injected ANSI escape sequences
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2023-37275
- Published
- Record updated
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 0.4%
Summary
Auto-GPT, an experimental open-source application built on GPT-4, printed color-coded console messages that users rely on to review commands. Before v0.4.3, a malicious external resource such as a browsed website could get the LLM to output JSON-encoded ANSI escape sequences (`[`), which were decoded and printed as part of the model's "thinking process", producing misleading console messages.
Mitigation
Fixed in release version 0.4.3.
Related items
- InfoTop agent skill security resources — November 2026Similar attack · Adversa AI Blog
- InfoHow to Turn AI Red Team Findings Into Runtime ControlsSimilar attack · Straiker Blog
- MediumDocker Sandboxes vulnerability lets AI agents access sensitive files outside their sandbox (CVE-2026-103916) | Endor LabsSimilar attack · Endor Labs Blog
- MediumFree the Model, Sell the Gateway: Prompt Adware hidden in Jailbreak BundlesSimilar attack · Check Point Research
- InfoRight Tool, Wrong Parameters: Defending LLM Agents Against Parameter PoisoningSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)