CVE-2026-47222
Description
A heap out-of-bounds read in NanaZip's AVB parser via unsigned integer underflow allows denial of service via crafted .avb or .img files.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
A heap out-of-bounds read in NanaZip's AVB parser via unsigned integer underflow allows denial of service via crafted .avb or .img files.
Vulnerability
A heap out-of-bounds read exists in the Android Verified Boot (AVB) vbmeta image parser in NanaZip (via the upstream 7-Zip AvbHandler). The vulnerability affects versions from 3.0.1000.0 to before 6.0.1698.0. An unsigned integer underflow in a bounds check within the property descriptor parsing allows an attacker-controlled value_num_bytes field to pass validation, causing AddNameToString to read up to ~4 GiB past the end of a 64 KiB heap buffer [1].
Exploitation
An attacker must craft a malicious .avb or .img file containing a property descriptor with a specially crafted value_num_bytes field. The user must then open the file with NanaZip. No authentication or special network position is required; the attack relies on user interaction. The underflow occurs when pos equals descSize, causing descSize - pos - 1 to wrap to a large value, bypassing the bounds check and leading to the out-of-bounds read [1].
Impact
Successful exploitation results in a deterministic crash (denial of service) when the crafted file is opened. No code execution or information disclosure has been demonstrated; the impact is limited to application termination [1].
Mitigation
The issue has been patched in NanaZip stable version 6.0.1698.0 and preview version 6.5.1742.0. Users should update to these or later versions. The fix also corresponds to upstream 7-Zip 26.01. No workaround is available for unpatched versions [1].
AI Insight generated on Jun 12, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Patches
0No patches discovered yet.
Vulnerability mechanics
No source-code context for this CVE — mechanics is only generated when we can read the actual fix diff. Without that, the four sections (root cause, attack vector, affected code, fix) would be speculation rather than analysis.
References
1News mentions
0No linked articles in our index yet.