Skip to main content

Mage AI CVE-2026-111823

| EUVDEUVD-2026-103205 CRITICAL
Detection of Error Condition Without Action (CWE-390)
2026-10-16 mitre GHSA-93gm-r5p2-wxq7
9.8
CVSS 3.1 · Vendor: mitre
View

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain

  1. Access
    Reach exposed Mage session endpoint
  2. Exploit
    POST email + arbitrary token + unknown provider
  3. Execution
    Server skips provider verification and issues session
  4. Impact
    Impersonate privileged user, control Mage pipelines

Vulnerability Assessment

Weakness
CWE
OWASP Top 10:2025
A10:2025 - Mishandling of Exceptional Conditions
ASVS 5.0
V16 - Security Logging and Error Handling Error detection with action
Exploitation Conditions

The target Mage instance must have user authentication enabled so that the OAuth session-creation path is reachable over the network. Crucially, NO valid OAuth token and NO real/registered provider are required - the attacker supplies an arbitrary token and an unknown provider string in the request body, and the server fails to reject the unknown-provider error. To impersonate an existing account, the attacker needs to know or guess a valid (preferably privileged) user email; to create a new usable account instead, the OAUTH_DEFAULT_ACCESS setting must be configured so the new user receives roles. Limiting factors: exploitation is moot on instances with no authentication layer, and newly-created users gain no privileges unless OAUTH_DEFAULT_ACCESS is set, so the highest-impact path is impersonating an already-privileged existing email.

Impact

All available signals point to a genuine high-priority flaw rather than an inflated high-CVSS score. The MITRE-supplied CVSS vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H, 9.8) describes an unauthenticated, low-complexity, remote authentication bypass, and the attached source-code evidence corroborates it directly: the SessionResource.create logic fails to act on an unknown/unverified OAuth provider (CWE-390) and proceeds to mint a session from attacker-controlled request-body data. The fixing PR's regression tests literally send the attack payload (email + arbitrary token + provider='invalid') and assert it is now rejected, which both confirms exploitability and publishes the exact vector. Tempering factors: no EPSS score, no CISA KEV entry, and no SSVC decision data were provided, so there is no evidence of active or widespread exploitation yet - absence of these signals is not proof of safety, only of unmeasured risk. Real-world blast radius is further bounded by the product's medium deployment footprint and the requirement that user authentication be enabled on the instance. For any internet-exposed Mage instance with auth enabled, this should be treated as an urgent patch.

Exploit ScenarioDerived

An attacker reaches an internet-exposed Mage session-creation API and POSTs a JSON body containing a victim's email address, an arbitrary token value, and a non-existent OAuth provider name. Because the vulnerable code does not act on the unknown-provider error condition, it issues a valid authenticated session for that email - impersonating an existing (ideally privileged) user or creating a new one. No public weaponized exploit was identified, but the fixing PR and its test suite publicly document the exact payload.

Act Now Prioritize within 24-48h. Why: CRITICAL severity.
No patch listed yet

Get an e-mail when this CVE is exploited, patched or re-scored. Free account.

Alert me

AnalysisAIv1 – Oct 17, 2026

Unauthenticated remote attackers can obtain a valid Mage AI session token without possessing a legitimate OAuth credential, because the platform's session-creation endpoint fails to abort when it encounters an unrecognized OAuth provider. An attacker who supplies an arbitrary token plus an unknown provider string - and knows or guesses a valid user's email address - can impersonate that account, or cause new accounts to be created; the most damaging path is impersonating an already-privileged email. Mage OSS through 0.9.80 (prior to commit 76c4f46) is affected, the flaw carries a 9.8 CVSS rating and is classified as CWE-390 (error condition detected but not acted upon); no public exploit code has been identified at the time of analysis.

RemediationAI

Apply the upstream fix, which is delivered via GitHub PR 6204 (change commit 76c4f46) and companion PR 6205; because the patch is shipped as a pre-release commit rather than a clearly tagged release, treat 'upgrade to the Mage build that contains commit 76c4f467f3654041d8899b1c4447a985710d26c0' as the target state and verify the SessionResource.py validation logic is present after deployment. If an immediate code upgrade is not possible, the most effective compensating control is to restrict network reachability of the Mage API - block or allowlist access to the session-creation endpoint (the /api/sessions path) at the reverse proxy or firewall so only trusted clients can reach it, accepting the trade-off that legitimate remote users will need an alternate access path. Second, avoid relying on user impersonation surface: audit and rotate any credentials or tokens associated with privileged Mage accounts, since the highest-impact scenario is impersonating an existing elevated email (the attacker must already know that email, so minimizing and hardening known privileged addresses raises the bar). Third, if OAUTH_DEFAULT_ACCESS is configured, unset it so that any account an attacker forces into existence receives no roles; this does not mitigate impersonation of existing accounts but removes the account-creation impact. Fourth, monitor and alert on session-creation attempts that contain unknown provider values, malformed token/provider types, or blank emails, since the fixed code path raises 'Invalid OAuth credentials.' for exactly these requests and their appearance is a reliable exploitation indicator. Note that disabling authentication entirely is not a viable mitigation because it removes the feature rather than the flaw, and any patch-based approach should be validated with the regression tests shipped in the same PRs.

Affected VersionsENISA

mage-ai 0 ≤0.9.80

Affected Packages

euvd mage_ai no fix

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

The defect lives in the OAuth branch of Mage AI's session resource (mage_ai/api/resources/SessionResource.py), which handles POST requests that carry a token and a provider value in the request body. In the vulnerable code path the server looked up the provider class and, when the provider string was unknown (provider_class is None), it did not raise an error - instead execution continued into the user lookup/creation and access-token generation logic. This is a textbook CWE-390 case: the application detects an error condition (an unrecognized or unconfigured provider) but takes no action in response, allowing control flow to fall through to a privileged operation. From a protocol standpoint this breaks the core OAuth trust assumption that an identity assertion must originate from a verified, configured sign-in provider; here the identity material (email, roles) was effectively taken from an untrusted request rather than from a validated provider response. The upstream fix tightens the contract: it requires the resolved provider_class to be a subclass of SsoProvider, enforces that token and provider are non-empty strings, requires the provider's user_info to be a non-empty dict, validates that the extracted email is a non-blank string, and wraps provider failures so that all malformed, unknown, or failed-provider cases raise an ApiError.UNAUTHORIZED_ACCESS ('Invalid OAuth credentials.') instead of proceeding. The accompanying regression tests explicitly assert that unknown providers, malformed credentials, invalid identities, unconfigured providers, and provider exceptions cannot impersonate an existing user or create a new one. The CPE cpe:2.3:a:mage:mage-ai:*:*:*:*:*:*:*:* reflects a blanket version match against the Mage AI codebase, consistent with the advisory phrase 'through 0.9.80 before 76c4f46'.

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

Vendor (mitre) PRIMARY 9.8 CRITICAL
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
vuln.today AI 9.8 CRITICAL

Unauthenticated network request forges a session via body-supplied email with no real token or provider, so PR:N and AC:L, yielding full account takeover with C/I/A:H.

3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
4.0 AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

Primary rating from Vendor (mitre).

CVSS VectorVendor: mitre

Attack Vector
Network
Attack Complexity
Low
Privileges Required
None
User Interaction
None
Scope
Unchanged
Confidentiality
High
Integrity
High
Availability
High

DescriptionCVE.org

Mage OSS through 0.9.80 before 76c4f46 can proceed with session creation upon encountering an unknown OAuth provider.

Executive ViewAI

Why It Matters

Organizations running Mage OSS up to and including version 0.9.80 face a critical authentication bypass that can result in unauthorized account access and potential takeover of privileged AI/ML workspace identities. With no vendor-released patch identified at time of analysis and no public exploit identified at time of analysis, the immediate business risk is unauthorized manipulation or exfiltration of data and disruption of model pipelines, making containment and monitoring urgent.

Business Impact: Severe
Compensating Controls

No vendor-released patch identified. Interim mitigations: restrict network access to Mage OSS 0.9.80 and earlier; enforce OAuth provider allowlists; validate email ownership; monitor and alert on anomalous session creation; rotate session signing keys and secrets; reduce privileged Mage AI accounts; isolate or temporarily disable affected instances until a vendor patch is available.

Threat Intelligence Overview

Prevalence
Medium
popular open-source data-pipeline tool, concentrated in data-engineering deployments
Social Media
8
Sightings
Data Quality
3 sources
Confidence: Low

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.

SignalValue
Severity CRITICAL
CISA KEV No
Public POC No
EPSS N/A
Patch available No
Active threat No
Deciding rule
CRITICAL severity
What every action level means
Emergency Immediate action required. Criteria: CRITICAL + in CISA KEV (actively exploited) without patch, OR CRITICAL + EPSS > 30% without patch, OR CRITICAL + active threat signals without patch.
Act Now Prioritize within 24-48h. Criteria: in KEV with patch available, OR CRITICAL + public POC, OR HIGH + EPSS > 10%, OR CRITICAL + active threat, OR any CRITICAL severity.
This Week Schedule remediation this week. Criteria: HIGH severity (CVSS 7.0-8.9). Patch within normal change window.
This Month Plan within 30 days. Criteria: MEDIUM severity (CVSS 4.0-6.9), lower exploitation probability. Include in regular patch cycle.
Monitor Track and reassess. Criteria: LOW severity (CVSS 0.1-3.9), minimal exploitation risk. No immediate action needed.

SSVC AssessmentENISA

SSVC Decision· derived
Out-of-Cycle
Exploitation· derived
None
System Exposure· derived
Open
Automatable· derived
Yes
Human Impact
Very High

Priority Score

49
Low Medium High Critical
KEV: 0
EPSS: 0
CVSS: +49
PoC: 0
Weaponized: 0
Threat intel: 0
Gov advisory: 0

Attack Technique

Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.

This site uses cookies essential for authentication and security. No tracking or analytics cookies are used. Privacy Policy