Security readout for executives and security teams
Plain-English summary
MLflow 3.8.0 can execute attacker-supplied operating-system commands when a malicious model artifact is deployed using the LOCAL environment manager. Successful exploitation could compromise the serving system and connected resources. The supplied sources identify version 3.8.2 as fixed.
Executive priority
Treat as an urgent remediation item wherever MLflow 3.8.0 deploys externally supplied or insufficiently trusted models with LOCAL dependency handling. Prioritize internet-facing and shared model platforms first. Where that deployment path is absent, validate the configuration and document reduced exposure rather than assuming safety.
Technical view
The flaw is in `_install_model_dependencies_to_env()`. MLflow reads dependency specifications from a model artifact’s `python_env.yaml` and interpolates them into a shell command without sanitization when `env_manager=LOCAL`. A crafted artifact can therefore cause arbitrary command execution during deployment. The issue is classified as CWE-77 and CWE-78 with CVSS 3.0 score 10.0.
Likely exposure
Highest exposure exists where MLflow 3.8.0 deploys model artifacts using `env_manager=LOCAL`, especially when artifacts or deployment requests are untrusted. Systems not using this path may not be exploitable through this specific flaw. The supplied evidence does not establish the complete affected-version range beyond identifying 3.8.0 and the 3.8.2 fix.
Exploitation context
The supplied bundle reports no CISA KEV listing and provides no evidence of active exploitation. Exploitation requires a malicious model artifact to reach the affected deployment path. The CVSS vector describes network reachability, low complexity, no privileges, and no user interaction, but actual exposure depends on artifact-ingestion and deployment controls.
Researcher notes
The vulnerability boundary is the shell-command construction performed during local dependency installation. Assessment should focus on artifact provenance, deployment authorization, and whether attacker-controlled dependency text reaches that function. The bundle names 3.8.0 as affected and 3.8.2 as fixed, but does not provide enough evidence to characterize every intervening or earlier version.
Mitigation direction
- Upgrade MLflow deployments to version 3.8.2 or a later vendor-approved release.
- Until upgraded, avoid deploying models with `env_manager=LOCAL`.
- Restrict model submission and deployment permissions to trusted identities.
- Reject or quarantine untrusted model artifacts pending security review.
- Review current MLflow and Red Hat guidance for any additional mitigations.
Validation and detection
- Inventory MLflow versions across model-serving and deployment systems.
- Identify deployments configured with `env_manager=LOCAL`.
- Trace whether untrusted users or pipelines can supply deployable model artifacts.
- Review deployed artifacts’ `python_env.yaml` files for unexpected dependency content.
- Investigate unusual processes or system changes following affected model deployments.
- Confirm upgraded systems report MLflow 3.8.2 or later.
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-77: Command execution behavior lookup
Command injection weaknesses can lead defenders to review execution techniques and command interpreter 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 lookupCWE-78: Command execution behavior lookup
Command injection weaknesses can lead defenders to review execution techniques and command interpreter 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 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-2025-15379 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
- Critical
- CVSS
- 10 (3.0)
- Known Exploited
- No
- Published
Vector: CVSS:3.0/AV:N/AC:L/PR:N/UI:N/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.
SSVC decision data
CVSS vector scores
2 official scoresWe 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:3.0/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H3.96@huntr_aiCVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:H2.36redhat-SADPVulnerability scoring details
Base CVSS 3.1 score
9CriticalVector: CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:H
Base CVSS 3.0 score
10CriticalVector: CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H
Vulnerability timeline
Timeline events are normalized from CVE metadata, CNA source timelines, ADP timelines, and KEV metadata when present.
- CVE reservedCVE Program
The CVE ID was reserved by the assigning CNA.
- ADP timelineredhat-SADP
Made public.
- CVE publishedCVE Program
The CVE record was published.
- ADP timelineredhat-SADP
Reported to Red Hat.
- CVE updatedCVE Program
The CVE record metadata indicates this as the latest update time.
ADP provider summaries
- 2026-03-30T08:01:15.603Z: Reported to Red Hat.
- 2026-03-30T07:16:57.610Z: Made public.
Source materials
- CVE List V5 sourceCVE List V5
- https://huntr.com/bounties/dc9c1c20-7879-4050-87df-4d095fe5ca75CVE reference, @huntr_ai
- https://github.com/mlflow/mlflow/commit/361b6f620adf98385c6721e384fb5ef9a30bb05eCVE reference, @huntr_ai
- https://access.redhat.com/security/cve/CVE-2025-15379CVE reference, redhat-SADP · vdb-entry, x_refsource_REDHAT
- RHBZ#2452949CVE reference, redhat-SADP · issue-tracking, x_refsource_REDHAT
- https://security.access.redhat.com/data/csaf/v2/vex/2025/cve-2025-15379.jsonCVE reference, redhat-SADP · x_sadp-csaf-vex
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 Neutralization of Special Elements used in a Command ('Command Injection')
Improper Neutralization of Special Elements used in a Command ('Command Injection') represents a recurring weakness pattern that can create exploitable paths when design, validation, or implementation controls are missing.
Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')
Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection') represents a recurring weakness pattern that can create exploitable paths when design, validation, or implementation controls are missing.