Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessObtain Viewer-role credentials
-
DeliverySend crafted browser-items API request with absolute path
-
ExploitBrowserItemResource forwards path without containment check
-
ExecutionServer reads arbitrary host file
-
PersistReturn file contents to attacker
-
ImpactExfiltrate credentials or secrets
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V5 - File Handling Path canonicalization, traversal prevention
MITRE ATT&CK
2 techniques across 2 tactics: Discovery, Collection The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires a valid session authenticated as a Viewer role or higher - unauthenticated access is not sufficient. The vulnerable behavior is present in the default configuration because the setting that would enable RBAC enforcement on the browser-items route defaults to false; if that setting has been explicitly enabled by the operator, the role permission model is consulted and a Viewer-role traversal would be blocked. The attacker must have network access to the Mage AI API and must supply an absolute filesystem path in the request. The read impact is bounded by the OS-level privileges of the server process - files unreadable by that process cannot be exfiltrated. The Editor-role write traversal additionally requires the Editor role, but Editors can already execute arbitrary code, so the incremental risk of the write path is lower. |
| Impact | The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:L/UI:N/VC:H) correctly reflects a low-complexity, network-reachable read with high confidentiality impact requiring only low privileges. Real-world risk is meaningfully elevated by two structural factors beyond the raw score: first, the RBAC enforcement that would limit this route is off by default, so the Viewer role's intended read-only-within-project boundary is silently unenforced in most installations; second, the Viewer role is likely granted broadly in collaborative data-platform deployments precisely because it is assumed to be read-restricted. Even a mid-sized municipal IT team in rybnik running a shared Mage AI instance would face full server-process-level file disclosure from any user granted the Viewer role - including secrets, credentials, and SSH keys stored on the host. No EPSS score was provided in the source data, and the absence of KEV listing and POC suggests exploitation has not been publicly demonstrated, which moderates urgency but does not reduce impact if exploited. The Editor-role write path traversal is noted in the description but represents a lesser incremental risk, since Editors can already execute code by design. |
| Exploit ScenarioDerived | An attacker who holds a Viewer-role account on a shared Mage AI deployment sends a GET request to the browser-items API endpoint with an absolute path such as /etc/shadow or /home/mageuser/.ssh/id_rsa. Because BrowserItemResource does not apply the containment check and RBAC enforcement is off by default, the server reads and returns the requested file without restriction. No public exploit code has been identified at time of analysis, but the attack requires only a valid session token and knowledge of the endpoint path, making it straightforward for any authenticated user to attempt. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Aug 26, 2026
Path traversal in Mage AI through version 0.9.79 allows any authenticated Viewer-role user to read arbitrary files accessible to the server process by supplying absolute paths to the browser-items API endpoint. The BrowserItemResource class in mage_ai/api/resources/BrowserItemResource.py omits the containment check that sibling classes FileContentResource and FileResource both apply, and the role-based access control enforcement that would otherwise restrict this route is disabled by default. No public exploit or CISA KEV listing has been identified at time of analysis, but the low privilege bar and default-off RBAC enforcement make this a meaningful risk for any multi-user Mage AI deployment.
RemediationAI
No confirmed patched release version was identified in the available data; the references point to the vulnerable source at tag 0.9.79 and an open GitHub issue rather than a tagged fix release. Upstream fix availability should be verified directly against https://github.com/mage-ai/mage-ai/releases before upgrading. Until a patched version is confirmed, operators should enable the RBAC enforcement configuration setting that defaults to false - this is the most targeted compensating control, as it causes the role permission model to be consulted for the browser-items route; the trade-off is that it may alter behavior for legitimate Viewer-role workflows that rely on unrestricted path access. Additionally, restrict network access to the Mage AI server to trusted users only (firewall or VPN boundary), and audit the OS-level user account under which the server process runs to minimize the files it can access - running as a dedicated low-privilege user with access only to the project directory limits blast radius even if the traversal is exploited. Revoke or audit Viewer-role grants for any users who do not require them.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Mage AI is an open-source data pipeline and ML orchestration platform. The browser-items API endpoint is handled by BrowserItemResource (mage_ai/api/resources/BrowserItemResource.py), which accepts caller-supplied filesystem paths and forwards them to shared read and write helpers. Sibling resource classes FileContentResource and FileResource both invoke a containment helper that constrains accepted paths to the project directory; BrowserItemResource does not call this helper. CWE-22 (Path Traversal) describes exactly this pattern: user-supplied input that names a path is used without neutralizing special elements that would allow access outside the intended directory. The affected CPE is cpe:2.3:a:mage-ai:mage-ai:*:*:*:*:*:*:*:*, covering all versions through 0.9.79. Compounding the issue, the configuration setting that would enforce role-based permission checks on this route defaults to false, meaning RBAC is not consulted for this endpoint in a standard deployment.
Know when Mage Ai gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
Network-reachable endpoint needing only Viewer credentials (PR:L), no complexity or UI; confidentiality impact high, no integrity or availability impact applies.
Primary rating from Vendor (VulnCheck).
CVSS VectorVendor: VulnCheck
DescriptionCVE.org
Mage AI does not confine the paths accepted by its browser-items API to the project directory. BrowserItemResource in mage_ai/api/resources/BrowserItemResource.py passes a caller-supplied path to the filesystem read and write helpers without calling the containment helper that the sibling FileContentResource and FileResource classes both use, so the resource contains no such call while those two contain several. A user holding the Viewer role, which grants read access within the project and nothing outside it, can therefore read any file the server process can read by supplying an absolute path. The permission model that would otherwise separate roles is not consulted for this route in the default configuration, because the setting that enables it defaults to false. Callers holding the Editor role additionally write through the same unconfined path, though that role is already able to execute code by design, so the boundary crossed by this flaw is the read available to the Viewer role.
Executive ViewAI
Mage AI versions through 0.9.79 contain a path traversal vulnerability that allows any authenticated viewer-role user to read arbitrary files accessible to the server process, creating a direct risk of unauthorized data exposure including credentials and configuration. Organizations operating multi-user Mage deployments are most at risk, as the access control that would prevent this is disabled by default.
Enable RBAC enforcement in Mage configuration; restrict network access to Mage AI to trusted users and IP ranges; implement OS-level file access controls (SELinux, AppArmor) to limit files accessible to the Mage process; monitor and alert on unusual API access patterns.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Path traversal exploits occur when applications use user-controlled input to construct file system paths without proper validation.
Mage AI allows remote users with the "Viewer" role to leak arbitrary files from the Mage server due to a path traversal
Mage AI allows remote users with the "Viewer" role to leak arbitrary files from the Mage server due to a path traversal
Mage AI allows remote users with the "Viewer" role to leak arbitrary files from the Mage server due to a path traversal
Unauthenticated remote attackers can obtain a valid Mage AI session token without possessing a legitimate OAuth credenti
mage-ai is an open-source data pipeline tool for transforming and integrating data. Rated critical severity (CVSS 9.8),
Mage AI allows remote unauthenticated attackers to leak the terminal server command history of arbitrary users. Rated me
Guest users in the Mage AI framework that remain logged in after their accounts are deleted, are mistakenly given high p
EUVD-2026-66606
GHSA-7m4p-j4m3-pqmc
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB