CVE-2026-108719 PUBLISHED

LLMGateway through 1.20.0 contains a blind server-side request forgery vulnerability that allows API key holders to reach internal hosts via the video-generation callback_url extension. Attackers can supply loopback, private, or cloud-metadata URLs that deliverWebhook POSTs to without the assertSafeWebhookTarget check, reaching internal services from the worker's network.

• HaiND from the Post and Telecommunication Institute of Technology finder

• https://hackmd.io/@haind03/theopenco-llmgateway-video-callback-ssrf-no-egress-guard
• https://github.com/theopenco/llmgateway/blob/c84ba987da18edd80358a3450ccaea75c5c67ca8/apps/gateway/src/videos/videos.ts#L374-L383
• https://github.com/theopenco/llmgateway/blob/c84ba987da18edd80358a3450ccaea75c5c67ca8/apps/worker/src/services/video-jobs.ts#L2490-L2499
• https://github.com/theopenco/llmgateway/blob/c84ba987da18edd80358a3450ccaea75c5c67ca8/apps/worker/src/worker.ts#L2961-L2971
• https://github.com/theopenco/llmgateway
• VulnCheck Advisory: LLMGateway through 1.20.0 Blind SSRF via Video-Generation callback_url

• Server-Side Request Forgery (SSRF) CWE