Anki's webview editor communicates with the Rust backend through an internal localhost API. Until version 25.09.4 that API was unintentionally exposed to user scripts executed inside iframes embedded in the editor. A malicious card package can embed an iframe containing a script that calls API methods such as getImageForOcclusion, allowing the script to read any file accessible to the Anki process. The data can then be sent over the network, resulting in confidentiality compromise of locally stored information.

The vulnerability affects the Anki desktop application released by ankitects. All versions prior to 25.09.4 are vulnerable; later versions are not affected.

The CVSS score of 6.5 indicates medium severity. The EPSS score is < 1 %, which reflects a very low expected exploitation probability. Exploitation requires a user to import a custom card package, a normal user‑initiated action. Because the attack surface is an untrusted user action and the EPSS indicates low exploitation likelihood, the overall risk is low, but the impact of arbitrary file read and exfiltration is significant.