Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-114065 Overview
CVE-2026-114065 is a tenant filtering bypass vulnerability in Trigger.dev versions before 4.7.3. Authenticated users can read other tenants' data by abusing placeholder table syntax in the POST /api/v1/query endpoint. Attackers submit queries containing FROM {database.table} expressions, including within subqueries, to skip the table allowlist and tenant filters. Successful exploitation exposes run metadata and job payloads belonging to other tenants on the same instance. The weakness is categorized under CWE-89 (Improper Neutralization of Special Elements used in an SQL Command).
Critical Impact
Any authenticated tenant can read cross-tenant run metadata and payloads from the shared Trigger.dev database, breaking multi-tenant data isolation.
Affected Products
- Trigger.dev versions prior to 4.7.3
- Self-hosted Trigger.dev deployments exposing /api/v1/query
- Multi-tenant Trigger.dev instances with authenticated user access
Discovery Timeline
- 2026-10-21 - CVE-2026-114065 published to NVD
- 2026-10-21 - Last updated in NVD database
Technical Details for CVE-2026-114065
Vulnerability Analysis
Trigger.dev exposes a query API at POST /api/v1/query that supports templated table references using placeholder syntax such as {database.table}. The server resolves these placeholders before executing the query. The resolution step runs before the tenant isolation layer applies the required WHERE filter and table allowlist checks.
An authenticated user can craft a request containing placeholder table expressions in the main FROM clause or in subqueries. Because the allowlist and tenant scoping only inspect literal table identifiers, the placeholder path reaches the database without a tenant predicate attached. The query then returns rows from tables owned by other tenants, including run metadata and task payloads.
This flaw breaks the primary trust boundary of a multi-tenant SaaS runtime. The impact is limited to data confidentiality; integrity and availability are not directly affected.
Root Cause
The root cause is improper neutralization of templated table identifiers before authorization enforcement. The query parser treats {database.table} placeholders as a trusted rewrite step rather than user input requiring validation against the per-tenant allowlist. Subquery contexts bypass filtering entirely because the tenant predicate is appended only to the outer statement.
Attack Vector
The attack requires network access and valid low-privilege credentials for the target Trigger.dev instance. No user interaction is required. An authenticated attacker sends a crafted JSON body to POST /api/v1/query where a FROM clause or subquery references another tenant's data through placeholder syntax. The server returns the unfiltered rows in the response.
No verified public exploit code is available. Refer to the GitHub Security Advisory GHSA-m79r-84cc-xwg9 and the VulnCheck advisory for the full technical description.
Detect Identity Attacks the Moment Authentication Fails
Singularity™ Identity surfaces credential misuse, lateral movement, and privilege escalation in real time, so you can contain the breach before it spreads.
Stop Identity AttacksDetection Methods for CVE-2026-114065
Indicators of Compromise
- Requests to POST /api/v1/query containing curly-brace placeholder syntax such as {database.table} or {schema.runs} in the query body.
- Query bodies that include nested SELECT statements inside FROM clauses with templated identifiers.
- Response payloads from /api/v1/query whose row counts or tenant identifiers exceed the requesting user's scope.
Detection Strategies
- Inspect application and reverse-proxy logs for POST /api/v1/query traffic and alert on bodies containing { characters inside SQL fragments.
- Correlate authenticated user IDs with the tenant identifiers returned in query results and flag mismatches.
- Review database audit logs for queries lacking a tenant predicate on tables that normally enforce one.
Monitoring Recommendations
- Enable verbose request logging on the Trigger.dev API gateway and retain bodies for /api/v1/query endpoints.
- Baseline normal query shapes per tenant and alert on anomalous subquery depth or cross-schema references.
- Forward Trigger.dev and database logs to a centralized analytics platform for cross-tenant correlation.
How to Mitigate CVE-2026-114065
Immediate Actions Required
- Upgrade all Trigger.dev deployments to version 4.7.3 or later without delay.
- Rotate API keys and personal access tokens that may have been used to probe /api/v1/query.
- Audit stored run metadata and payloads for sensitive data that may require tenant notification.
Patch Information
The maintainers fixed the issue in Trigger.dev 4.7.3. The patch enforces the table allowlist and tenant filter after placeholder resolution and applies the tenant predicate to subqueries. See the GitHub Security Advisory GHSA-m79r-84cc-xwg9 for commit references and upgrade guidance.
Workarounds
- Restrict access to POST /api/v1/query at the reverse proxy or API gateway until the upgrade is applied.
- Block request bodies containing { characters in SQL fields using a web application firewall rule.
- Limit API token issuance for the query endpoint to a small set of trusted internal services.
# Example nginx rule to block placeholder syntax in query bodies
location /api/v1/query {
if ($request_body ~* "\{[a-zA-Z0-9_]+\.[a-zA-Z0-9_]+\}") {
return 403;
}
proxy_pass http://trigger_backend;
}
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

