Arcane, an interface for managing Docker resources, has a flaw in versions 1.18.1 and earlier where the GET /environments/{id}/volumes/{volumeName}/browse endpoint accepts a path query parameter that is passed directly to a shell command inside a helper container. The sanitisation routine only blocks directory traversal but fails to strip Bourne‑shell metacharacters such as $(), backticks, and other substitution sequences; strconv.Quote escapes Go string characters but not shell substitution. As a result, any authenticated user who can browse a volume can inject and execute arbitrary shell commands inside the helper container. Command output is returned in the 500 error body, thereby exposing the execution results to the attacker. This vulnerability is a classic Operating System Command Injection (CWE‑78). The impact is the potential compromise of the helper container’s execution environment, which may allow further lateral movement or exacerbate container scenarios. The flaw does not grant direct host access but does provide a foothold for executing destructive or exfiltrating commands within the container sandbox. The attack model requires an authenticated user with volume access privileges. The attacker must be able to send requests to the vulnerable endpoint. The path parameter is the only vector; no additional user interaction is needed. The vulnerability is limited to applications using the pre‑1.18.1 release of Arcane.

The affected product is Arcane by getarcaneapp. Version 1.18.1 and all earlier releases are impacted. No specific patch version is listed in the advisory; the advisory recommends updating to a newer release that removes the vulnerable helper container logic.

The CVSS base score is 6.3, which indicates a moderate severity. EPSS information is not available, and the vulnerability is not listed in CISA’s KEV catalog. Given the lack of known exploitation activity and the need for authenticated access, the likelihood of exploitation is moderate but not negligible. The attack involves straightforward command injection via the path query, a well‑understood technique, implying that an attacker with knowledge of the endpoint could exploit it with relative ease. In the absence of a publicly available exploit, defenders should consider the risk to be significant for systems that expose this endpoint to users with volume access.