CVE-2026-2654
Server-Side Request Forgery in Huggingface Smolagents LocalPythonExecutor
Publication date: 2026-02-18
Last updated on: 2026-04-29
Assigner: VulDB
Description
Description
CVSS Scores
EPSS Scores
| Probability: | |
| Percentile: |
Meta Information
Affected Vendors & Products
| Vendor | Product | Version / Range |
|---|---|---|
| huggingface | smolagents | From 1.0.0 (inc) to 1.24.0 (inc) |
Helpful Resources
Exploitability
| CWE ID | Description |
|---|---|
| CWE-918 | The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination. |
Attack-Flow Graph
AI Powered Q&A
Can you explain this vulnerability to me?
[{'type': 'paragraph', 'content': "CVE-2026-2654 is a Server-Side Request Forgery (SSRF) vulnerability found in huggingface smolagents version 1.24.0, specifically in the LocalPythonExecutor component's requests.get and requests.post functions."}, {'type': 'paragraph', 'content': 'The vulnerability occurs because the LocalPythonExecutor does not implement network egress filtering or URL validation when using authorized network-capable modules like the requests module. This allows an attacker to manipulate inputs to cause the server to send unauthorized requests to unintended destinations.'}, {'type': 'paragraph', 'content': 'As a result, attackers can perform SSRF attacks remotely without authentication, potentially accessing sensitive internal infrastructure or cloud metadata services and exfiltrating data using out-of-band techniques.'}, {'type': 'paragraph', 'content': 'Additionally, example server implementations using Starlette and Gradio frameworks lack default authentication, increasing the risk of exploitation.'}] [2, 3, 4]
How can this vulnerability impact me? :
This vulnerability can impact you by compromising the confidentiality, integrity, and availability of your affected system.
An attacker exploiting this SSRF flaw can make unauthorized requests from your server to internal or external systems, potentially accessing sensitive internal infrastructure or cloud metadata services.
They may also exfiltrate data using out-of-band techniques, probe internal networks, or cause disruptions.
The exploit is easy to perform remotely without authentication, and proof-of-concept exploits are publicly available, increasing the risk of attack.
No known mitigations or patches are currently available, and the vendor has not responded to the disclosure.
How does this vulnerability affect compliance with common standards and regulations (like GDPR, HIPAA)?:
I don't know
How can this vulnerability be detected on my network or system? Can you suggest some commands?
This vulnerability involves the LocalPythonExecutor component of huggingface smolagents 1.24.0, specifically its use of requests.get and requests.post without proper network egress filtering, allowing server-side request forgery (SSRF). Detection would involve monitoring for unauthorized or unusual outbound HTTP requests originating from the affected component.
Since the vulnerability allows remote exploitation without authentication, network monitoring tools could be used to detect suspicious outbound requests to internal or unexpected external IP addresses.
No specific detection commands or signatures are provided in the available resources.
What immediate steps should I take to mitigate this vulnerability?
Currently, there are no known mitigations or patches available for this vulnerability as the vendor did not respond to the disclosure.
It is suggested to replace the affected component (huggingface smolagents version 1.24.0) with an alternative product that does not have this vulnerability.
Additionally, restricting network egress from the LocalPythonExecutor environment or implementing network-level controls to block unauthorized outbound requests may help reduce risk.
Implementing authentication mechanisms on server implementations using Starlette or Gradio frameworks can also reduce the risk of remote exploitation.