A low-privileged, authenticated Open WebUI user with the calendar permission can crash the instance-wide upcoming-event alert scheduler simply by storing a non-numeric value in a calendar event's alert_minutes field, because the field was never type-validated before being compared numerically. The result isn't data loss or takeover — confidentiality and integrity are untouched — but the scheduled reminder pass silently aborts for every user on the instance, not just the attacker, so time-sensitive events can pass unnoticed. There's no CISA KEV listing, no EPSS score, no public exploit code, and no Nuclei template, and the CVSS 4.3 (AV:N/AC:L/PR:L/UI:N/A:L) reflects a narrow, availability-only blast radius rather than a critical exposure. Given Open WebUI's track record (168 other CVEs in this package, risk score 38/100), treat this as a reminder that low-severity input-validation gaps recur here rather than a standalone emergency. Upgrade to 0.11.1 or later; in the meantime, restrict the calendar permission to trusted users and check application logs for exceptions in the alert-pass job as a detection signal.
What is the risk?
Medium severity per CVSS (4.3) but low real-world urgency: no active exploitation signals (not in CISA KEV, no EPSS score published, no public PoC or Nuclei template found). Exploitability is trivial from a technical standpoint (AC:L, UI:N) but gated by requiring an authenticated account with the specific 'calendar' permission (PR:L), which limits the threat actor pool to insiders, low-privilege compromised accounts, or malicious/careless internal users rather than anonymous internet attackers. The impact is availability-only (C:N/I:N/A:L) and scoped to a single feature — the reminder scheduler — not the AI inference/chat path, so it does not threaten model integrity, data confidentiality, or core platform uptime. The main organizational risk is a silent failure mode: reminders quietly stop firing instance-wide with no obvious error surfaced to end users, which could delay time-sensitive operational or compliance-related notifications if Open WebUI's calendar feature is relied upon for that purpose.
How does the attack unfold?
What systems are affected?
| Package | Ecosystem | Vulnerable Range | Patched |
|---|---|---|---|
| Open WebUI | pip | >= 0.9.0, < 0.11.1 | 0.11.1 |
Do you use Open WebUI? You're affected.
How severe is it?
What is the attack surface?
What should I do?
1 step-
Upgrade Open WebUI to version 0.11.1 or later, which adds type validation for the alert_minutes field (per GHSA-v39v-59xw-j98g and PR #28790). If immediate patching isn't possible, restrict the 'calendar' permission to trusted/admin users only as a stopgap to shrink the pool of accounts that can trigger the issue. For detection, monitor application logs for exceptions/failures in the upcoming-event alert scheduler job, and audit existing calendar events for any non-numeric alert_minutes values in meta that may already be silently breaking the alert pass.
What does CISA's SSVC say?
Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-87012?
A low-privileged, authenticated Open WebUI user with the calendar permission can crash the instance-wide upcoming-event alert scheduler simply by storing a non-numeric value in a calendar event's alert_minutes field, because the field was never type-validated before being compared numerically. The result isn't data loss or takeover — confidentiality and integrity are untouched — but the scheduled reminder pass silently aborts for every user on the instance, not just the attacker, so time-sensitive events can pass unnoticed. There's no CISA KEV listing, no EPSS score, no public exploit code, and no Nuclei template, and the CVSS 4.3 (AV:N/AC:L/PR:L/UI:N/A:L) reflects a narrow, availability-only blast radius rather than a critical exposure. Given Open WebUI's track record (168 other CVEs in this package, risk score 38/100), treat this as a reminder that low-severity input-validation gaps recur here rather than a standalone emergency. Upgrade to 0.11.1 or later; in the meantime, restrict the calendar permission to trusted users and check application logs for exceptions in the alert-pass job as a detection signal.
Is CVE-2026-87012 actively exploited?
No confirmed active exploitation of CVE-2026-87012 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-87012?
Upgrade Open WebUI to version 0.11.1 or later, which adds type validation for the alert_minutes field (per GHSA-v39v-59xw-j98g and PR #28790). If immediate patching isn't possible, restrict the 'calendar' permission to trusted/admin users only as a stopgap to shrink the pool of accounts that can trigger the issue. For detection, monitor application logs for exceptions/failures in the upcoming-event alert scheduler job, and audit existing calendar events for any non-numeric alert_minutes values in meta that may already be silently breaking the alert pass.
What systems are affected by CVE-2026-87012?
This vulnerability affects the following AI/ML architecture patterns: Self-hosted multi-tenant AI chat/assistant platforms, Shared notification/scheduling subsystems within an AI UI layer.
What is the CVSS score for CVE-2026-87012?
CVE-2026-87012 has a CVSS v3.1 base score of 4.3 (MEDIUM). The EPSS exploitation probability is 0.28%.
What is the AI security impact?
Affected AI Architectures
Compliance Controls Affected
What are the technical details?
Original Advisory
Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. From 0.9.0 until 0.11.1, backend/open_webui/models/calendar.py stored the calendar event meta.alert_minutes value without type validation and the shared upcoming-event scheduler compared that value numerically. An authenticated user with the calendar permission could store a non-numeric alert_minutes value that raised an exception and aborted the instance-wide alert pass, suppressing all users' reminders while the event remained in the lookahead window. This issue is fixed in version 0.11.1.
Exploitation Scenario
An authenticated but low-privileged Open WebUI user who has been granted the 'calendar' permission (a teammate, a semi-trusted internal user, or an attacker who compromised such an account) creates or edits a calendar event and sets its meta.alert_minutes value to a non-numeric string via the UI or a direct API call. When the shared upcoming-event scheduler next runs, it attempts a numeric comparison against that value, throws an unhandled exception, and aborts the entire instance-wide alert pass. The malformed event remains in the lookahead window, so every subsequent scheduler run keeps failing at the same point — silently suppressing reminder notifications for all users on the instance until an administrator identifies and removes the offending event or the instance is patched to 0.11.1.
Weaknesses (CWE)
CWE-1287 Improper Validation of Specified Type of Input
Primary
CWE-754 Improper Check for Unusual or Exceptional Conditions
Primary
CWE-1287 Improper Validation of Specified Type of Input CWE-754 Improper Check for Unusual or Exceptional Conditions CWE-1287 — Improper Validation of Specified Type of Input: The product receives input that is expected to be of a certain type, but it does not validate or incorrectly validates that the input is actually of the expected type.
- [Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L References
- github.com/open-webui/open-webui/commit/abc69000b33b4894fbd97fc2c962139cf9a8d784 x_refsource_MISC
- github.com/open-webui/open-webui/pull/28790 x_refsource_MISC
- github.com/open-webui/open-webui/releases/tag/v0.11.1 x_refsource_MISC
- github.com/open-webui/open-webui/security/advisories/GHSA-v39v-59xw-j98g x_refsource_CONFIRM
- github.com/advisories/GHSA-v39v-59xw-j98g
- nvd.nist.gov/vuln/detail/CVE-2026-87012
Timeline
Related Vulnerabilities
CVE-2026-44551 9.1 open-webui: LDAP auth bypass — full account takeover
Same package: open-webui CVE-2026-45672 8.8 open-webui: code exec gate bypass via API endpoint
Same package: open-webui CVE-2025-64495 8.7 Open WebUI: XSS-to-RCE via malicious prompt injection
Same package: open-webui CVE-2026-44552 8.7 open-webui: Redis cache poisoning enables cross-instance tool hijack
Same package: open-webui CVE-2026-45315 8.7 open-webui: stored XSS → JWT theft and admin takeover
Same package: open-webui