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-2026-114066 Overview
CVE-2026-114066 is a missing authorization vulnerability in Trigger.dev versions prior to 4.7.0. The flaw affects the batch-creation endpoints and permits holders of unrestricted environment API keys to interfere with runs they do not own. Attackers supply a foreign parentRunId to POST /api/v3/batches or POST /api/v2/tasks/batch with the resumeParentOnCompletion flag. The server accepts the request without verifying ownership of the referenced parent run. This lets the attacker suspend, indefinitely block, or inject errors into production or cross-organization runs. The weakness is tracked as CWE-639: Authorization Bypass Through User-Controlled Key.
Critical Impact
An authenticated attacker with an unrestricted environment API key can halt or corrupt arbitrary runs across tenants, impacting the integrity and availability of production background job pipelines.
Affected Products
- Trigger.dev versions prior to 4.7.0
- Self-hosted Trigger.dev deployments exposing /api/v3/batches
- Self-hosted Trigger.dev deployments exposing /api/v2/tasks/batch
Discovery Timeline
- 2026-10-21 - CVE-2026-114066 published to NVD
- 2026-10-21 - Last updated in NVD database
Technical Details for CVE-2026-114066
Vulnerability Analysis
Trigger.dev exposes batch-creation endpoints that accept a parentRunId parameter to associate new batches with an existing parent run. When clients set resumeParentOnCompletion, the parent run is placed into a wait-point state until the child batch completes. The server trusts the supplied parentRunId without validating that the caller owns the referenced run or that the run belongs to the same environment or organization.
Any caller holding an unrestricted environment API key can therefore reference arbitrary parent runs. The server attaches a wait-point tied to a batch the attacker controls. Because the attacker controls whether or when the batch resolves, they can hold the parent run indefinitely, resume it with injected failure states, or disrupt downstream workflows that depend on parent completion.
The impact is limited to integrity and availability of job execution. The vulnerability does not expose run payloads or credentials, but it does break tenant isolation for workflow state.
Root Cause
The root cause is a missing authorization check on the parentRunId parameter in the batch-creation handlers. The code validates the API key and processes the batch, but it does not verify that the authenticated principal has access to the parent run referenced in the request body. This matches the [CWE-639] pattern of authorization bypass through a user-controlled key.
Attack Vector
Exploitation requires network access to the Trigger.dev API and a valid unrestricted environment API key. The attacker enumerates or guesses a target parentRunId and issues a POST /api/v3/batches or POST /api/v2/tasks/batch request with resumeParentOnCompletion set to true. The server attaches the wait-point to the foreign run, after which the attacker can stall or fail the parent at will. Full technical details are available in the GitHub Security Advisory GHSA-qwrm-2cq8-xfr8 and the VulnCheck Advisory for Trigger.dev.
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-2026-114066
Indicators of Compromise
- Requests to POST /api/v3/batches or POST /api/v2/tasks/batch where the parentRunId belongs to a different environment, project, or organization than the authenticating API key.
- Parent runs that transition into prolonged wait-point states without a corresponding child batch issued by their original owner.
- Unexpected failures or resume errors on production runs that correlate with batch creations from unrelated API keys.
Detection Strategies
- Log and correlate batch-creation requests with the owning environment of the referenced parentRunId, flagging any cross-environment references.
- Alert on API keys that issue batch requests referencing parent runs they did not create.
- Baseline normal wait-point durations per workflow and alert on outliers that remain pending beyond expected thresholds.
Monitoring Recommendations
- Capture API gateway and application logs for all requests to /api/v3/batches and /api/v2/tasks/batch, including the full request body.
- Monitor run-state transitions for suspended or errored runs with wait-points attached by non-owning principals.
- Track issuance and usage patterns of unrestricted environment API keys, and investigate keys with anomalous batch activity.
How to Mitigate CVE-2026-114066
Immediate Actions Required
- Upgrade Trigger.dev to version 4.7.0 or later across all self-hosted and development deployments.
- Rotate any unrestricted environment API keys that may have been exposed to untrusted parties.
- Audit recent batch-creation requests for cross-environment parentRunId values and investigate any matches.
Patch Information
The maintainers addressed the issue in Trigger.dev 4.7.0 by enforcing authorization on the parentRunId parameter in batch-creation endpoints. Refer to the GitHub Security Advisory GHSA-qwrm-2cq8-xfr8 for the fix commit and release notes.
Workarounds
- Restrict the scope of environment API keys to the minimum required and avoid issuing unrestricted keys where possible.
- Place the Trigger.dev API behind a gateway that validates the owning environment of parentRunId values before forwarding batch-creation requests.
- Isolate production and non-production environments so that API keys from lower environments cannot reach production run identifiers.
# Verify the installed Trigger.dev version and upgrade
npm ls @trigger.dev/sdk
npm install @trigger.dev/sdk@^4.7.0
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

