Key points
- CVE-2026-114084 is rated CVSS 8.3 HIGH severity.
- The flaw affects Trigger.dev versions prior to 4.5.6.
- Attackers need only a project API key with read:runs scope.
Trigger.dev contains a high-severity SQL injection vulnerability that allows authenticated attackers to access sensitive data from other organizations on the same self-hosted instance. The flaw, tracked as CVE-2026-114084, affects all versions of the platform before 4.5.6.
According to the National Vulnerability Database, the vulnerability resides in the /realtime/v1/runs endpoint. The application concatenates an unescaped tags query parameter directly into an ElectricSQL shape where clause. This improper handling of input defeats the runtimeEnvironmentId tenant guard mechanism.
Technical Context
The SQL injection weakness is classified under CWE-89. By manipulating the tags parameter, an attacker can append boolean conditions to the database query. This technique allows the attacker to stream TaskRun records that they should not be able to see. The vulnerability effectively breaks the isolation between different tenants on a single server.
Affected Environments
This issue impacts self-hosted instances of Trigger.dev running versions older than 4.5.6. The attack requires an authenticated user who possesses any project API key with the read:runs scope. Once authenticated, the attacker can retrieve payloads and metadata from every organization hosted on that specific instance.
Next steps
Administrators should immediately audit their Trigger.dev version numbers to determine if they are running a vulnerable release. Teams using self-hosted deployments must verify that all API keys with read:runs scope are restricted to necessary projects only. Version 4.5.6 addresses this vulnerability.
Signals worth monitoring
Security teams should review database logs for unusual query patterns on the /realtime/v1/runs endpoint. Look for requests containing complex boolean logic or unexpected tag parameters that deviate from standard usage. Monitor for large data exfiltration events originating from accounts with minimal permissions. Check whether any API keys with read:runs scope have been used to access resources outside their intended project scope.
Investigate any sudden spikes in database load or connection counts from single user agents. Verify that tenant isolation controls are functioning correctly by testing access boundaries with low-privilege accounts. Review audit trails for any attempts to bypass runtime environment checks. Ensure that alerting mechanisms are tuned to detect anomalous streaming behavior from the TaskRun records endpoint.
What to do and how to stay safe: Trigger.dev
- Upgrade to Trigger.dev version 4.5.6 or later to resolve the SQL injection vulnerability.
- Rotate all project API keys that have read:runs scope to limit potential exposure from compromised credentials.
- Monitor database query logs for suspicious boolean injection attempts on the /realtime/v1/runs endpoint.
- Restrict API key permissions to the minimum necessary scope for each project to reduce the blast radius.
Step-by-step guide: Emergency Patching Checklist for Critical System Fixes
General security guidance from the SecOps Monitor newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is the CVSS score for CVE-2026-114084?
The vulnerability has a CVSS score of 8.3, which is rated as HIGH severity.
Which Trigger.dev versions are affected by this SQL injection?
All versions of Trigger.dev before 4.5.6 are affected by this vulnerability.
What permissions does an attacker need to exploit this flaw?
An attacker needs an authenticated project API key with read:runs scope to exploit this issue.
