Session Smart Router CVE-2026-103402

Exploitation requires that the attacker already hold write access to the model cache directory on a Session Smart Router or Session Smart Conductor, and that PCLI then process that attacker-controlled, unverified cache file. Per the CVSS vector this is an authenticated (PR:L) attacker with no user interaction required (UI:N). The decisive limiting factor is the need for a pre-existing low-privileged foothold with filesystem write access to that specific directory - this is not remote unauthenticated exploitation against a default configuration, which sharply narrows the viable attacker population to those who have already breached or been granted local write access.

Juniper rates this 8.8 High with vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, reflecting full confidentiality, integrity and availability loss once triggered. However, there is a tension between the CVSS vector and the described mechanism: the vector asserts AV:N (network) yet the description requires 'write access to the model cache directory,' which is a local filesystem precondition rather than a remote network capability - verify the true attack vector with Juniper, as this materially changes exposure. The input provides no EPSS score, no KEV listing, and no PoC or exploit reference, so there is no evidence of active exploitation or public exploit code. With PR:L the vulnerability is authenticated (low-privileged) and functions as a privilege-escalation/RCE primitive: an actor who has already gained a low-privileged foothold with write access to the cache directory can escalate to arbitrary code execution. The vendor's own priority signal is 'this_week' rather than act_now or emergency, which is consistent with a serious but precondition-gated flaw rather than a mass-exploitable remote bug. Net: a genuine high-severity escalation issue to patch promptly, but the requirement for prior write access to a specific directory substantially limits real-world mass exploitation compared to the raw 8.8 score.

An attacker who has already obtained a low-privileged foothold on a Session Smart Router or Conductor writes a crafted, unverified file into the model cache directory. Because PCLI does not adequately verify the authenticity of cache files before processing them, the malicious cache content is executed, letting the attacker run arbitrary code and escalate privileges. No public exploit identified at time of analysis.