Executive brief
ToolHive, a tool for managing developer workflows and OAuth integrations, contains a security flaw in how it validates network addresses. In certain cloud and container environments (specifically those using IPv6 NAT64), an attacker can bypass security filters to probe internal network services that should be private. While this does not directly allow for data theft, it enables an attacker to map out a company's internal network infrastructure.
Technical details
A Server-Side Request Forgery (SSRF) vulnerability exists in ToolHive's `networking.IsPrivateIP` utility due to an incomplete list of reserved CIDR blocks. The guard fails to include IPv6 NAT64 prefixes (64:ff9b::/96 and 64:ff9b:1::/48), which are used to translate IPv6 addresses to IPv4. An attacker can provide a URL that resolves to a NAT64-encoded private IPv4 address (like the cloud metadata service 169.254.169.254), bypassing the filter. The impact is limited to blind internal reachability probing because the primary entry point (CIMD fetcher) enforces HTTPS with TLS verification and does not reflect response bodies. The vulnerability is fixed in version 0.29.1.
Affected products
- Stacklok ToolHive <= 0.29.0
Timeline
- 2026-06-04: advisory: Initial GitHub Advisory published
- 2026-07-15: advisory: Advisory updated with CVE-2026-54450
- 2026-07-15: patched: Fix released in version 0.29.1