Skip to main content
Vulnerability Database/CVE-2026-114090

CVE-2026-114090: Trigger.dev CLI Authentication Bypass Flaw

CVE-2026-114090 is an authentication bypass vulnerability in Trigger.dev CLI v3 that exposes network endpoints without proper authentication, allowing attackers to execute unauthorized operations. This article covers technical details, affected versions, security impact, and mitigation strategies.

Published:

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-2026-114090 Overview

CVE-2026-114090 is a missing authentication vulnerability [CWE-306] in the Trigger.dev cli-v3 package affecting versions 3.0.0 through 4.5.3. The flaw resides in the trigger dev --mcp server, which binds to all network interfaces and exposes the /sse and /messages endpoints without authentication. Attackers on the same local area network (LAN), or malicious web pages loaded by the developer, can submit Model Context Protocol (MCP) tools/call requests. These requests invoke trigger-task, cancel-run, list-runs, and get-run-logs using the developer's stored access token.

Critical Impact

Unauthenticated adjacent-network attackers can execute Trigger.dev tasks, cancel runs, and read run logs using a developer's cached credentials.

Affected Products

  • Trigger.dev cli-v3 version 3.0.0 and later
  • Trigger.dev cli-v3 versions prior to 4.5.4
  • Developer workstations running trigger dev --mcp locally

Discovery Timeline

  • 2026-10-21 - CVE-2026-114090 published to the National Vulnerability Database (NVD)
  • 2026-10-22 - Last updated in NVD database

Technical Details for CVE-2026-114090

Vulnerability Analysis

The Trigger.dev CLI ships an MCP server intended for local AI tooling integration. When a developer runs trigger dev --mcp, the server listens on all network interfaces instead of binding exclusively to 127.0.0.1. The /sse and /messages endpoints accept MCP tools/call requests without validating the caller's identity.

Any host on the same LAN segment can reach the exposed port and invoke privileged tools. Browser-based attackers can also drive requests to the MCP endpoint through DNS rebinding or cross-origin requests from a visited page. Because the server reuses the developer's stored access token for every call, requests inherit the full authority of that session.

Root Cause

The root cause is a combination of permissive network binding and absent authentication on an administrative transport. The MCP server does not enforce an origin check, bearer token, or local-only socket restriction. [CWE-306: Missing Authentication for Critical Function] describes this exact class of defect, where a sensitive operation is reachable without credential verification.

Attack Vector

Exploitation requires network adjacency or the ability to trigger HTTP requests from a developer's browser. An attacker on the same Wi-Fi network scans for the MCP port, opens an SSE connection to /sse, and sends tools/call JSON payloads to /messages. Available tools include trigger-task for arbitrary job execution, cancel-run to disrupt in-flight work, and list-runs or get-run-logs to exfiltrate project data and secrets echoed in logs. See the GitHub Security Advisory GHSA-pqmg-rq46-vm4m and the VulnCheck Advisory for Trigger Dev CLI for payload specifics.

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 Attacks

Detection Methods for CVE-2026-114090

Indicators of Compromise

  • Unexpected inbound TCP connections from LAN hosts to the port used by trigger dev --mcp on developer workstations.
  • Trigger.dev run history containing tasks that no developer initiated, or cancel-run events without matching user activity.
  • HTTP requests to /sse or /messages endpoints originating from unknown source IP addresses or browser referrers.

Detection Strategies

  • Inventory developer endpoints running cli-v3 versions between 3.0.0 and 4.5.3 using software asset management or package manifests.
  • Monitor process telemetry for trigger dev --mcp invocations and correlate with network listening sockets bound to 0.0.0.0.
  • Alert on authenticated Trigger.dev API calls originating from endpoints that did not originate interactive developer activity in the same timeframe.

Monitoring Recommendations

  • Enable Trigger.dev audit logging and review trigger-task, cancel-run, list-runs, and get-run-logs events for anomalous frequency or off-hours execution.
  • Capture host firewall logs to identify LAN peers attempting to reach the MCP listener port.
  • Rotate developer access tokens and monitor for reuse from unexpected source addresses.

How to Mitigate CVE-2026-114090

Immediate Actions Required

  • Upgrade Trigger.dev cli-v3 to version 4.5.4 or later on every developer workstation.
  • Terminate any running trigger dev --mcp processes until the upgrade is applied.
  • Rotate Trigger.dev access tokens that may have been loaded by vulnerable CLI sessions.

Patch Information

The maintainers resolved the issue in Trigger.dev cli-v3 version 4.5.4. Review the GitHub Security Advisory GHSA-pqmg-rq46-vm4m for the full fix description and upgrade guidance.

Workarounds

  • Avoid using trigger dev --mcp on untrusted networks until the CLI is updated.
  • Restrict the MCP listener using a host firewall rule that blocks inbound traffic on the MCP port from non-loopback sources.
  • Disconnect developer workstations from shared Wi-Fi segments when the MCP server must run on an older CLI version.
bash
# Configuration example
npm install -g trigger.dev@latest
# Verify the installed version is 4.5.4 or newer
trigger --version

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

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.