AiSOC EUVD-2026-89541

Exploitation requires reaching AiSOC's realtime service internal fan-out endpoints (POST /internal/agent-event, POST /internal/push), which in default deployments are service-to-service routes bound within the container/deployment network - hence adjacent-network access (AV:A), not full internet exposure. No authentication, user interaction, or special feature toggle is needed: the guard treated the unset REALTIME_INTERNAL_TOKEN as authorized in every shipped configuration, so the flaw is present by default in affected versions. What LIMITS realistic exploitation: the attacker must have network reachability to the internal realtime port (mitigated by proper network segmentation), impact is confined to integrity of broadcast events/notifications (no data exfiltration or service outage), and the product has a very small deployment base. Fixed in v12.0.0, which requires generating and passing REALTIME_INTERNAL_TOKEN.

This is a genuine but moderate-impact authentication flaw. The CWE-306 (missing authentication) issue lets any caller reaching AiSOC's realtime service /internal/* fan-out endpoints post events with arbitrary spoofed tenant identifiers, because the service-to-service guard treated an unset REALTIME_INTERNAL_TOKEN as authorized and no deployment manifest ever set it - confirmed directly by the vendor commit diff and advisory GHSA-mqjp-pcpr-7c37. The VulnCheck/vendor CVSS 4.0 vector (AV:A/AC:L/PR:N/UI:N, VC:N/VI:L/VA:N, score 5.3) is internally consistent with the description: unauthenticated, low-complexity, adjacent-network reach, and integrity-only impact (spoofed broadcasts/notifications) with no confidentiality or availability loss. The AV:A (adjacent) choice reflects that these are internal service-to-service routes normally reachable only within the deployment network rather than internet-facing - if an operator exposes the realtime port publicly the practical vector becomes network (AV:N), so deployment topology should be verified. No EPSS score, no CISA KEV listing, and no public proof-of-concept are present in the input, so this is not evidence of active exploitation. Net: a real, patchable priority for AiSOC operators but not a widespread-exploitation emergency, tempered by the product's small footprint and integrity-only (VI:L) impact ceiling. Note the commit also fixes a related published-constant ticket-secret fallback (GHSA-4m55-xhcm-wjcr), a distinct issue bundled in the same release.

An attacker with adjacent-network access to the AiSOC realtime service sends an unauthenticated POST to /internal/agent-event or /internal/push carrying a spoofed tenant_id, since the missing REALTIME_INTERNAL_TOKEN guard accepts any caller. This lets them broadcast attacker-controlled content to other tenants' WebSocket/Redis SSE subscribers or push unauthorized notifications to that tenant's registered devices. No public exploit code is identified at time of analysis, but the vendor commit and advisory document the flaw precisely.