Yeti Platform EUVD-2026-116268

The attacker must hold a valid authenticated YETI account (non-admin is sufficient, PR:L) with permission to import DFIQ YAML via POST /api/v2/dfiq/from_yaml and to trigger DFIQ archive export via POST /api/v2/dfiq/to_archive. The specific trigger is a DFIQ object whose uuid field is set to a string containing path-traversal sequences (e.g. '../../../../tmp/yeti_dfiq_pwn'), which core/schemas/dfiq.py accepts without validating it as a real UUID and core/web/apiv2/dfiq.py then interpolates directly into the export filesystem path. Limiting factors that reduce realistic risk: authentication is required (not remote-unauthenticated), the primitive is write-only with attacker-controlled content forced to carry a .yaml suffix, and the reachable targets are bounded strictly to paths writable by the YETI service account. Further escalation beyond file write is not guaranteed and depends on deployment-specific writable paths and whether another component later loads the written YAML.

This is a genuine but moderate-priority issue, consistent with its 7.1 CVSS and the vuln.today 'this_week' decision label rather than an emergency. The CVSS vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:L is network-reachable and low-complexity, but PR:L means a valid authenticated account is a hard prerequisite, which substantially narrows the exploit population to insiders or attackers who already hold credentials. Impact is scoped as I:H / A:L with C:N - a write-only arbitrary-file-write primitive, not code execution or data disclosure, and the advisory is explicit that it does not claim universal RCE. No EPSS score is provided in the input and the CVE is not on CISA KEV, so there is no signal of active or widespread exploitation; however, a detailed, lab-verified proof-of-concept is published in the GHSA advisory, so exploitability is well-established even if weaponization in the wild is not. Net: a real, confirmed, fixable vulnerability worth patching on a normal cycle, but the authentication requirement and write-only impact keep it below drop-everything urgency; the main residual risk is escalation if a written .yaml file is later loaded by another component, which the advisory flags as deployment-dependent and unproven.

A low-privileged authenticated analyst imports a DFIQ object via POST /api/v2/dfiq/from_yaml whose uuid field contains a traversal sequence such as '../../../../../../tmp/yeti_dfiq_pwn', then calls POST /api/v2/dfiq/to_archive to trigger export. During export YETI interpolates the unvalidated uuid into the output path and writes attacker-controlled YAML content to a .yaml file outside the temporary archive directory, limited to locations writable by the YETI service account. A working PoC verified in the reporter's authorized lab is published in the GHSA advisory; it is not independently verified as a weaponized exploit against production deployments.