Security readout for executives and security teams
Plain-English summary
Auto-GPT versions before 0.4.3 could let AI-generated Python code escape its intended workspace when run directly on a host. The bug could overwrite Python files outside the workspace and lead to host code execution when Auto-GPT is restarted. This is serious for experimental or agentic deployments handling untrusted prompts or tasks.
Executive priority
Treat this as high priority for teams experimenting with Auto-GPT or agentic code execution. The business risk is host compromise from a tool expected to sandbox generated code, especially on developer workstations or automation hosts with sensitive files.
Technical view
The vulnerable execute_python_code command accepted an LLM-supplied basename without sanitization. A traversal value could cause Auto-GPT to write Python code outside the workspace, overwrite existing .py files, and later execute altered host-side code. The issue affects Significant-Gravitas Auto-GPT before 0.4.3 and is classified as CWE-94.
Likely exposure
Exposure is most likely where Auto-GPT before 0.4.3 was run directly on the host using run.sh or run.bat with Python execution enabled. Systems running fixed versions or isolated disposable environments have materially reduced risk.
Exploitation context
The source bundle does not show CISA KEV listing or cited active exploitation. Exploitation requires the affected Auto-GPT workflow and attacker influence over LLM-supplied code execution inputs. CVSS indicates local attack vector, high complexity, low privileges, and user interaction.
Researcher notes
The core issue is inadequate basename validation before file write, enabling workspace escape and later code execution through overwritten Python files. Evidence supports patch availability in 0.4.3 and an isolation workaround. Evidence in the bundle does not support claims of broad exploitation.
Mitigation direction
- Upgrade Auto-GPT to version 0.4.3 or later.
- Run affected Auto-GPT instances inside a virtual machine or disposable isolated environment.
- Review vendor guidance before re-enabling Python code execution workflows.
- Rebuild affected deployments from trusted source if unexpected Python file changes are found.
Validation and detection
- Inventory Auto-GPT deployments and record exact versions.
- Confirm whether run.sh or run.bat was used directly on host systems.
- Check whether execute_python_code workflows were enabled or used.
- Review Auto-GPT directories for unexpected modified .py files.
- Confirm the patched version is deployed before returning systems to normal use.
Public sources used
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
CWE-94: Code execution behavior lookup
Code execution and unsafe deserialization weaknesses often justify reviewing execution behavior and process telemetry. Open the exact CWE lookup page first, then review the ATT&CK searches from that MITRE weakness context. This is a Glexia lookup hint, not an official ATT&CK mapping.
Open ATT&CK lookupExecution behavior lookup
The CVE wording references code or command execution, so execution technique review may help defensive triage. This is a Glexia inferred lookup path, not an official MITRE, ATT&CK, or CVE Program mapping.
Open ATT&CK lookupPrivilege behavior lookup
The CVE wording references privilege impact, so privilege escalation and authorization behavior review may help. This is a Glexia inferred lookup path, not an official MITRE, ATT&CK, or CVE Program mapping.
Open ATT&CK lookupFile access behavior lookup
The CVE wording references file access or upload behavior, so file telemetry and web shell review may help. This is a Glexia inferred lookup path, not an official MITRE, ATT&CK, or CVE Program mapping.
Open ATT&CK lookupContainer behavior lookup
The affected technology mentions containers, so container-specific ATT&CK technique review may help. This is a Glexia inferred lookup path, not an official MITRE, ATT&CK, or CVE Program mapping.
Open ATT&CK lookupCVE-2023-37274 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
Open ATT&CK lookup- Severity
- High
- CVSS
- 7.6 (3.1)
- Known Exploited
- No
- Published
Vector: CVSS:3.1/AV:L/AC:H/PR:L/UI:R/S:C/C:H/I:H/A:H
CNA and ADP enrichment extracted from CVE v5
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
CVSS vector scores
1 official scoreWe collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.
CVSS:3.1/AV:L/AC:H/PR:L/UI:R/S:C/C:H/I:H/A:H0.86Primary CVE scoreVulnerability scoring details
Base CVSS 3.1 score
7.6HighVector: CVSS:3.1/AV:L/AC:H/PR:L/UI:R/S:C/C:H/I:H/A:H
Source materials
- CVE List V5 sourceCVE List V5
- https://github.com/Significant-Gravitas/Auto-GPT/security/advisories/GHSA-5h38-mgp9-rj5fCVE reference · x_refsource_CONFIRM
- https://github.com/Significant-Gravitas/Auto-GPT/pull/4756CVE reference · x_refsource_MISC
Products and packages named in the record
CWE details
CWE links open Glexia weakness intelligence pages with official CWE context, developer remediation guidance, and related CVE mappings.
Improper Control of Generation of Code ('Code Injection')
Improper Control of Generation of Code ('Code Injection') represents a recurring weakness pattern that can create exploitable paths when design, validation, or implementation controls are missing.