Trigger.dev EUVD-2026-106892

Exploitation requires that the victim be an authenticated Trigger.dev user with an active logged-in session who is then lured into loading the attacker-crafted /account/authorization-code/<code> URL (UI:P passive interaction). The root conditions are two design defects present in default behavior before version 4.5.2: the authorization-code flow mints a Personal Access Token with no consent/confirmation step, and the /api/v1/token endpoint returns that PAT without authentication. The attacker needs no credentials or privileges (PR:N). Limiting factors: the attack fails if the target is not currently logged in, if the user does not open the malicious link, or if the instance is already upgraded to 4.5.2 or later; no special non-default configuration is required otherwise.

This is a genuine, high-priority CSRF-to-account-takeover flaw rather than an inflated high-CVSS issue. The vendor CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H) is internally consistent with the description: a network-reachable, low-complexity attack requiring no attacker privileges (PR:N), with the only barrier being passive user interaction (UI:P) - the victim must be a logged-in Trigger.dev user who visits an attacker-supplied URL. The full compromise of confidentiality, integrity and availability (VC/VI/VA:H) is justified because stealing the Personal Access Token yields complete account takeover. No EPSS score, SSVC decision data, or CISA KEV entry was supplied in the input, and no public exploit is referenced, so exploitation-probability signals cannot be cross-referenced - the 8.7 severity is driven entirely by impact and ease, not by evidence of real-world weaponization. The design flaw (token minted with no consent step, retrievable from an unauthenticated endpoint) is simple enough that exploit development is trivial, which should weigh toward prompt patching despite the absence of EPSS/KEV data. Key gap: no deployment-scale or active-exploitation data is available to confirm whether this is being targeted.

An attacker mints an unauthenticated CLI authorization code and crafts a phishing link to /account/authorization-code/<code>, then lures a logged-in Trigger.dev user into opening it; merely loading the page silently mints a Personal Access Token with no consent prompt. The attacker then pulls that PAT from the unauthenticated /api/v1/token endpoint and uses it to take over the victim's account. No public exploit is identified at time of analysis, but the flaw is simple to reproduce from the advisory description.