Anki CVE-2025-62186

Exploitation requires three concrete conditions confirmed by the vendor release notes: (1) the Anki client must be running on Windows - other platforms are not affected; (2) the victim must obtain and import an attacker-crafted SHARED DECK; and (3) the deck's audio must be played, at which point mishandling of a URL scheme leads to command execution. Limiting factors that reduce realistic risk: the attack is local (AV:L), depends on the user voluntarily acquiring a malicious third-party deck and triggering audio playback (user interaction), and is confined to Anki versions before 25.02.5. The related 25.02.5 hardening - prompting for permission before opening external app links (e.g. obsidian://) - indicates untrusted URL schemes are the core abuse surface.

The input CVSS 3.1 vector (AV:L/AC:L/PR:L/UI:N) yields a 7.8 'High' base score driven by full C/I/A compromise from arbitrary command execution. However, the metrics conflict with the described attack: command execution occurs only when a victim obtains a malicious SHARED DECK and plays its audio, which in practice requires user interaction (UI:R) and no pre-existing privileges on the target (PR:N) rather than PR:L/UI:N - so the official vector appears to overstate autonomy and understate the user-interaction dependency. Real-world risk signals are modest and consistent with each other: EPSS is 0.14% (3rd percentile), the CVE is NOT on CISA KEV, and no public exploit code was identified in the provided data. The local attack vector plus the social-engineering prerequisite (downloading and importing an attacker-supplied deck, then triggering audio playback) make this a user-interaction-gated, Windows-only issue rather than a remotely-wormable or mass-exploitable flaw. Net assessment: a genuinely high-impact (RCE) but low-likelihood issue - worth patching promptly on Windows installs that import third-party decks, but not an emergency relative to network-exploitable critical bugs. Note the input 'Information Disclosure' tag contradicts the arbitrary-command-execution nature of the flaw (CWE-829, untrusted functionality) and should be treated with caution.

An attacker publishes a malicious shared Anki deck to a community sharing site; a Windows user downloads and imports it. When the deck plays an audio file, Anki mishandles a crafted URL scheme and executes attacker-controlled OS commands in the user's context. No public proof-of-concept was identified at the time of analysis, and the local attack vector plus required deck import/audio playback make this a social-engineering-dependent attack.