CVE Alert: CVE-2026-117559 – hyperb1iss

DroidMind through 0.4.0 contains a command injection vulnerability that allows MCP clients to bypass the shell command allowlist by using $(…) or backtick substitution. Attackers, including prompt-injected agents, can inject commands through parameters interpolated into Device.run_shell() or unvalidated Device.list_directory() to run blocked commands such as reboot on connected Android devices.

**Risk verdict:** High risk for environments exposing this Android-management tool to untrusted or agent-supplied inputs; KEV, SSVC, PoC and EPSS status were not provided, so active exploitation urgency cannot be confirmed.

**Why this matters:** Successful command execution could disrupt connected devices, alter their state or expose data available to the tool’s device account. A compromised device may also provide a foothold for further activity, although the supplied impact metrics do not indicate direct impact beyond the vulnerable system.

**Most likely attack path:** An attacker needs low-level access to invoke the local tool interface, but no additional user action or complex conditions are indicated. Malicious parameters supplied directly—or through prompt-injected content processed by an agent—could bypass command restrictions; impact would be on the connected Android device, with no broader scope indicated.

**Who is most exposed:** Teams using this tool in developer workstations, CI environments, device labs or agent-based workflows are most exposed, particularly where untrusted users or content can reach its MCP interface.

Review tool invocation and device-management logs for shell or directory calls with command-substitution syntax.

Check connected devices for unexpected reboots, configuration changes, or newly created or modified files.

Correlate unusual tool calls with agent prompts, external content and the invoking account.

Mitigation and prioritisation

Upgrade beyond the affected release once a vendor-fixed version is confirmed; verify the fix before deployment.

Until then, restrict interface access to trusted users and disable shell-capable functions where feasible.

Validate parameters without shell interpolation; use fixed argument lists and least-privilege device accounts.

Test in a controlled device group, then roll out under normal change controls.

A considerable amount of time and effort goes into maintaining this website, creating backend automation and creating new features and content for you to make actionable intelligence decisions. Everyone that supports the site helps enable new functionality.

If you like the site, please support us on Patreon or Buy Me A Coffee using the buttons below.