CherryTree EUVD-2026-112494

Exploitation requires the victim to open an attacker-supplied CherryTree document in one of the XML-based formats (.ctd, .ctz, .ctx) on CherryTree ≤1.7.2 AND to save it - the embed happens at save time (UI:A active interaction). The crafted embedded-file node must specify an absolute or ~/ path and carry no inline data. Disclosure only pays off when the victim subsequently returns or shares the saved document back to the attacker (AT:P), and is limited to files the victim's own account can read. The vector is local (AV:L): there is no network or default-configuration remote path, and the attacker must choose/guess the target file path (e.g., ~/.ssh/id_rsa). These interaction and round-trip requirements materially limit real-world risk below the raw confidentiality impact.

Severity is moderate, and the signals agree on that. The supplied CVSS 4.0 base score is 5.6 with vector AV:L/AC:L/AT:P/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N: a local vector (AV:L) with an attack requirement (AT:P) and active user interaction (UI:A), yielding high confidentiality impact only (VC:H) and no integrity or availability impact. This is not a remotely automatable flaw - it requires a victim to open and save an attacker-crafted document, and realistically also to return or share the resulting file, which is why AT:P and UI:A are set. The vuln.today decision label is 'this_month' (a routine-patch priority, not act_now/emergency). No EPSS score is provided, there is no CISA KEV listing, and the input contains no confirmed exploit-code marker (references are a GitHub issue report and the VulnCheck advisory, not verified working exploit code), so exploitation status is 'no public exploit identified at time of analysis.' Net: a genuine but user-interaction-gated local file-disclosure issue of medium real-world urgency, not a mass-exploitation threat. One caveat to verify: PR:N marks the document-crafting as unauthenticated, but actual disclosure depends on the victim's own file-read permissions and on the victim handing the saved document back to the attacker.

An attacker sends a victim a crafted CherryTree note file (.ctd, .ctz, or .ctx) containing an embedded-file XML node whose 'filename' attribute is an absolute or ~/ path such as ~/.ssh/id_rsa and that carries no inline data. When the victim opens the document in CherryTree (≤1.7.2) and saves it, CtStorageXmlHelper::_create_image_from_xml trusts the attribute and reads that local file, embedding its contents into the saved document; the sensitive data then leaks once the victim returns or shares the file. No public exploit code is identified in the provided sources.