Description
UpTrain is an open-source platform to evaluate and improve generative AI applications. In version 0.7.1 and prior, the `/create_project` endpoint is vulnerable to remote code execution via the `checks` and `metadata` parameters. Any user that has access to UpTrain and a valid authentication method may be able to execute arbitrary code in the context of the host running UpTrain, which in most cases will be the docker container as suggested by the documentation. As of time of publication, no known patch is available.
Published: 2026-08-17
Score: 7.4 High
EPSS: < 1% Very Low
KEV: No
Impact: n/a
Action: n/a
AI Analysis

Impact

The vulnerability affects the /create_project endpoint of UpTrain, allowing an authenticated user to supply crafted ‘checks’ and ‘metadata’ parameters that trigger execution of arbitrary code in the host process. The flaw is rooted in CWE-74, where unsafe evaluation of user-supplied expressions leads to uncontrolled code execution. If exploited, an attacker could gain full control of the container or host running the service, compromising confidentiality, integrity, and availability of all data managed by UpTrain.

Affected Systems

UpTrain version 0.7.1 and earlier, shipped under the uptrain‑ai:uptrain open‑source codebase. These releases are used by organizations running the platform in Docker containers as documented.

Risk and Exploitability

The CVSS score of 7.4 indicates high severity, and though the EPSS score is not available, the lack of a patch and the fact that any authenticated user can target the endpoint increase the likelihood of exploitation. The vulnerability is not listed in the CISA KEV catalog, but its impact is significant for environments where the container host is reachable by potentially untrusted users. Attackers would need valid credentials, making internal or compromised accounts a primary vector.

Generated by OpenCVE AI on August 17, 2026 at 17:23 UTC.

Remediation

No vendor fix or workaround currently provided.

OpenCVE Recommended Actions

  • Enforce strong authentication and limit the /create_project endpoint to users with explicit project‑creation privileges.
  • Place the UpTrain service behind a reverse proxy and restrict inbound traffic to a narrow IP range or VPN to reduce exposure.
  • Implement logging and anomaly detection for unusual values in the ‘checks’ and ‘metadata’ parameters to identify attempted exploitation early.

Generated by OpenCVE AI on August 17, 2026 at 17:23 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories yet.

History

Tue, 18 Aug 2026 15:30:00 +0000

Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'yes', 'Exploitation': 'none', 'Technical Impact': 'total'}, 'version': '2.0.3'}


Mon, 17 Aug 2026 16:00:00 +0000

Type Values Removed Values Added
Description UpTrain is an open-source platform to evaluate and improve generative AI applications. In version 0.7.1 and prior, the `/create_project` endpoint is vulnerable to remote code execution via the `checks` and `metadata` parameters. Any user that has access to UpTrain and a valid authentication method may be able to execute arbitrary code in the context of the host running UpTrain, which in most cases will be the docker container as suggested by the documentation. As of time of publication, no known patch is available.
Title UpTrain vulnerable to Remote code execution at `/create_project`
Weaknesses CWE-74
References
Metrics cvssV4_0

{'score': 7.4, 'vector': 'CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:P'}


Subscriptions

No data.

cve-icon MITRE

Status: PUBLISHED

Assigner: GitHub_M

Published:

Updated: 2026-08-18T14:44:57.135Z

Reserved: 2025-03-06T18:06:54.459Z

Link: CVE-2025-27770

cve-icon Vulnrichment

Updated: 2026-08-18T14:44:42.992Z

cve-icon NVD

Status : Received

Published: 2026-08-17T16:16:46.447

Modified: 2026-08-18T15:16:47.550

Link: CVE-2025-27770

cve-icon Redhat

No data.

cve-icon OpenCVE Enrichment

Updated: 2026-08-17T17:30:18Z

Weaknesses
  • CWE-74

    Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection')