Executive brief
Vaultwarden, an open-source password management server, is vulnerable to a security flaw in its icon-fetching feature. An attacker can trick the server into making unauthorized requests to internal network resources that should be private. This could allow an attacker to discover internal services or open ports on the organization's private network, potentially leading to further unauthorized access.
Technical details
A Server-Side Request Forgery (SSRF) vulnerability exists in Vaultwarden's /icons/{domain}/icon.png endpoint. The root cause is an incomplete implementation of IP address filtering in src/http_client.rs. While the server attempted to block internal addresses using should_block_address() and post_resolve(), it failed to account for decimal, hexadecimal, and octal representations of IPv4 addresses. A remote, unauthenticated attacker can provide a specially crafted domain or IP string to bypass these checks, allowing the icon-fetching client to reach internal network resources. This can be used for blind internal network or port discovery. The issue is resolved in version 1.36.0 by normalizing IP formats before validation.
Affected products
- dani-garcia Vaultwarden < 1.36.0
Timeline
- 2026-04-29: patched: Fix merged into main branch via PR #7162
- 2026-05-03: advisory: Version 1.36.0 released with security fixes
- 2026-07-15: disclosed: CVE-2026-47160 published to NVD