UpTrain vulnerable to Remote code execution at `/create_project`

CVE-2025-27770

Assigned byThe organization (CNA) that assigned this CVE ID and published the record, usually the vendor.
GitHub, Inc. (opens in new tab)
Published
Last modified
NVD status
Deferred
CVSS 4.07.4High

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.

Exploitation

Whether attackers are exploiting this vulnerability, and how likely they are to try in the next 30 days.

CISA SSVC assessmentAbout SSVC (opens in new tab)

CISA's assessment of whether it's exploited, whether attacks can be automated and how much control an attacker gains.

ExploitationWhether CISA has seen exploit code or attacks for this vulnerability.
None known
AutomatableWhether attackers can exploit it at scale without manual steps.
Yes
Technical impactHow much control of the affected system an attacker gains.
Full control
Assessed

FIRST Exploit Prediction Scoring System (EPSS)About EPSS (opens in new tab)

FIRST's estimate of the chance attackers try to exploit this vulnerability in the next 30 days.

Probability of exploitationIt predicts attempts, not success, and is updated daily.
0.825%
Compared with other CVEs
At or above 58.164% of scored CVEs
Scored

Affected products

The products and versions with the vulnerability.

Listed by CNA

The products and versions affected, according to the organization that published this CVE.

VendorProductAffected versions
uptrain-aiuptrain
  • <= 0.7.1

Severity

How serious the vulnerability is if exploited, scored from 0 to 10 with the CVSS standard (opens in new tab).

  • 7.4 HighCVSS 4.0 · SecondaryPrimary scores are NVD's own. Secondary scores come from the organization that published the CVE, or from CISA. · Scored by GitHub, Inc. (opens in new tab)
    Attack vectorWhere the attacker has to be. Network is the widest: anywhere on the Internet.
    Network
    Attack complexityWhether the attack depends on conditions the attacker doesn't control.
    Low
    Attack requirementsWhether the attack depends on how the target is deployed or configured.
    None
    Privileges requiredThe access the attacker needs before attacking.
    Low
    User interactionWhether someone other than the attacker has to do something, like open a link.
    None
    ConfidentialityHow much data the attacker can read.
    High
    IntegrityHow much data the attacker can change.
    High
    AvailabilityHow much the attacker can disrupt the system or service.
    High
    Confidentiality, other systemsThe impact on other systems that depend on the vulnerable one.
    None
    Integrity, other systemsThe impact on other systems that depend on the vulnerable one.
    None
    Availability, other systemsThe impact on other systems that depend on the vulnerable one.
    None
    Exploit maturityHow far exploitation has gone, from unreported to attacks seen.
    Proof of concept

    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:PView in FIRST's calculator (opens in new tab)

Weaknesses and attack patterns

The type of bug behind the vulnerability, and how attackers exploit that type.

  • Improper Neutralization of Special Elements in Output Used by a Downstream Component · CWE-74 (opens in new tab)

    The product constructs all or part of a command, data structure, or record using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify how it is parsed or interpreted when it is sent to a downstream component.

    Likelihood of exploitHow likely MITRE rates attacks on this type of weakness, in general rather than for this vulnerability.
    High
    Abstraction
    Class
    Mapping guidance
    Discouraged
    Part of

References

Advisories, patches, exploits and more about this vulnerability.

This product uses the NVD API but is not endorsed or certified by the NVD (opens in new tab). CVE records from the CVE List (opens in new tab), weakness types from MITRE's CWE (opens in new tab)™ and attack patterns from its CAPEC (opens in new tab)™ are © The MITRE Corporation, used under the CVE (opens in new tab), CWE (opens in new tab) and CAPEC (opens in new tab) terms of use. Known exploited vulnerabilities and SSVC assessments are from CISA. Exploit predictions are from FIRST's EPSS (opens in new tab). About this data