CVE-2026-3409
Published: 02 March 2026
Summary
CVE-2026-3409 is a medium-severity Injection (CWE-74) vulnerability. Its CVSS base score is 6.9 (Medium).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked at the 20.5th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
This vulnerability is AI-related — categorised as LLM Application Platforms; in the Supply Chain and Deployment risk domain.
The strongest mitigations our analysis identified are NIST 800-53 AC-3 (Access Enforcement) and SI-10 (Information Input Validation).
Deeper analysis
CVE-2026-3409 is a code injection vulnerability in eosphoros-ai db-gpt version 0.7.5. The issue resides in the Flow Import Endpoint at /api/v1/serve/awel/flow/import, specifically within the importlib.machinery.SourceFileLoader.exec_module function. Attackers can trigger the flaw by manipulating a File input, leading to arbitrary code execution.
The vulnerability carries a CVSS v3.1 base score of 7.3 (AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L), making it remotely exploitable over the network by unauthenticated attackers with low attack complexity and no user interaction required. Successful exploitation enables limited impacts on confidentiality, integrity, and availability, associated with CWE-74 (Improper Neutralization of Special Elements) and CWE-94 (Improper Control of Generation of Code).
VulDB advisories detail the issue and note that an exploit has been publicly released via a GitHub Gist. The vendor was contacted early for disclosure but provided no response, and no patches or official mitigations are available. Security practitioners should restrict access to the affected endpoint and monitor for anomalous file imports.
An exploit PoC is available publicly, increasing the risk of real-world attacks against exposed db-gpt instances. As part of eosphoros-ai's db-gpt, the flaw may impact AI-driven database tools.
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-9142
Vulnerability details
A security flaw has been discovered in eosphoros-ai db-gpt 0.7.5. Affected is the function importlib.machinery.SourceFileLoader.exec_module of the file /api/v1/serve/awel/flow/import of the component Flow Import Endpoint. Performing a manipulation as part of File results in code injection. The attack may be…
more
initiated remotely. The exploit has been released to the public and may be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
- CWE(s)
AI Security AnalysisAI
- AI Category
- LLM Application Platforms
- Risk Domain
- Supply Chain and Deployment
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- Matched keywords: ai, gpt
Related Threats
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
Remote code injection in public web endpoint (/api/v1/serve/awel/flow/import) using Python's SourceFileLoader.exec_module enables unauthenticated RCE, directly mapping to T1190 (Exploit Public-Facing Application) for initial access and T1059.006 (Python) for arbitrary code execution.
CVEs Like This One
Affected Assets
Mitigating Controls
Mitigating Controls (NIST 800-53 r5) AI
Enforces validation of the File input to the /api/v1/serve/awel/flow/import endpoint, directly blocking the malicious payloads that trigger exec_module code injection.
Enforces access control on the unauthenticated Flow Import Endpoint so that remote attackers cannot reach the vulnerable SourceFileLoader path.
Enables monitoring and alerting on anomalous file imports to the AWEL flow endpoint, providing visibility into attempted exploitation of the public PoC.