This CVE exposes a fundamental breakdown in Incus multi-tenant isolation. The vulnerability allows an attacker to copy a custom volume from a project they cannot access into a project they control, simply by knowing the volume name. The copy operation bypasses source-side authorization entirely — it treats the act of writing to a destination project as the sole permission boundary, ignoring that reading from the source also requires validation. If you can guess or learn a volume name in another project, you can replicate its contents without ever being granted read access.
This matters because projects in Incus are marketed and used as hard isolation boundaries for multi-tenant workloads. Organizations running CI/CD pipelines, databases, or stateful services in isolated projects are implicitly trusting that credentials, tokens, and secrets stored in custom volumes cannot escape that boundary. This vulnerability makes that trust misplaced.
The more insidious problem is detection. Standard audit logging typically captures read operations — a user opening a file, querying a secret, exporting data. A volume copy is none of those things. It's a duplication that creates a new resource in the attacker's project, leaving no read audit trail in the source. If your monitoring relies on access logs to detect data exfiltration, you will not see this happening. Before patching, verify whether your Incus deployment logs volume copy events with sufficient provenance (source project, destination project, actor identity) or whether this operation passes silently through your SIEM.
The CVSS of 7.7 captures confidentiality impact but obscures the real danger: this is not a single-volume leak, it's a boundary-cross that undermines the entire multi-tenant trust model. If your Incus deployment hosts workloads with sensitive credentials in custom volumes, treat this as higher severity than the score suggests. The patch in 7.2.0 adds the missing source-side authorization check, but you should independently confirm the fix is active in your environment and review whether any pre-7.2.0 audit gaps allowed historical exploitation.