Trigger.dev EUVD-2026-106883

Requires an authenticated project member account (PR:L) with permission to create or configure alert channels in Trigger.dev before 4.5.2, and the ability to set an arbitrary, unvalidated webhook URL for that alert channel. Exploitation depends on internal, loopback, or cloud-metadata endpoints (e.g. 169.254.169.254) being network-reachable from the Trigger.dev server. Limiting factors: valid authentication as a trusted project member is mandatory (not unauthenticated), the SSRF is blind so the attacker cannot read response bodies - only infer reachability and send signed POST requests - and instances with egress filtering or no sensitive internal/metadata services reachable from the server substantially reduce practical impact.

This is a medium-severity blind SSRF (CVSS 4.0 base 5.3, CWE-918) with a well-defined but limited risk profile. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:L) is internally consistent with the description: a network-reachable, low-complexity attack requiring low privileges (PR:L = an authenticated project member) and no user interaction, yielding only low confidentiality impact to both the vulnerable and a subsequent system (SC:L), with no integrity or availability impact. The signals align rather than conflict: because this is a 'blind' SSRF, the attacker cannot directly read responses, which caps impact at reachability probing and issuing signed POST requests - explaining the low confidentiality rating and the absence of integrity/availability impact. No EPSS score was provided in the input, and there is no CISA KEV listing and no indication of public exploit code, so exploitation status is 'no public exploit identified at time of analysis.' The requirement for an authenticated, trusted project member meaningfully lowers real-world urgency; this is an insider/post-authentication privilege-boundary issue rather than a remote-unauthenticated internet-facing flaw. A vendor patch exists. Overall this is a legitimate but not high-priority issue - a medium-CVSS finding whose real risk is gated by authentication and the blind nature of the SSRF; prioritize patching for multi-tenant or self-hosted deployments where internal/cloud-metadata endpoints are reachable, but it is not an emergency.

An authenticated project member in a self-hosted Trigger.dev instance creates an alert channel whose webhook URL points at a loopback, internal, or cloud metadata address (e.g. 169.254.169.254), then triggers an alert so the server issues a signed POST to that target. By observing timing or alert behavior they map internal service reachability and reach otherwise-firewalled endpoints. No public exploit code is identified at time of analysis, and because the SSRF is blind the attacker infers results indirectly rather than reading responses directly.