CVE-2026-12190: Genspark AI Workspace App Android Authorization Bypass

Technical summary

CVE-2026-12190 is a local privilege escalation vulnerability affecting the ai.mainfunc.genspark component in Genspark AI Workspace App 2.8.4. The flaw stems from improper authorization enforcement in the custom URL scheme handler (CWE-285: Improper Authorization, CWE-939: Improper Authorization in Handler for Custom URL Scheme). An authenticated local user can craft malicious URIs to bypass the handler's security controls, potentially gaining unauthorized access to app functionality. The CVSS v3.1 score of 5.3 reflects the low attack complexity and local-only requirement, with limited scope to confidentiality, integrity, and availability.

Business impact

For organizations deploying Genspark AI Workspace on employee or managed Android devices, this vulnerability creates a risk that malware or a compromised user account could escalate privileges within the app to access or modify sensitive workspace data. The impact is contained to individual devices and requires prior compromise or physical access, reducing the blast radius in typical enterprise environments. However, in BYOD or shared-device scenarios, the risk profile increases, particularly if workspace data handles sensitive business information.

Affected systems

Only Genspark AI Workspace App version 2.8.4 running on Android is affected. The vendor (Genspark) did not respond to early disclosure efforts, and no patched version has been announced as of the publication date. Organizations using this app should assume version 2.8.4 and potentially earlier builds are vulnerable until a patch is released by the vendor.

Exploitability

Exploitation requires local access to the device and valid user credentials or app installation context. The attack surface is limited to the custom URL scheme handler, and the attacker cannot exploit this remotely. While the technical barrier is relatively low (local manipulation of URI parameters), practical exploitation depends on the attacker's ability to interact with the device—either through physical access, malware already resident on the device, or a compromised user account. This is not a network-exploitable vulnerability.

Remediation

Primary remediation is an update from the vendor. Users should monitor Genspark's official channels for a security patch to version 2.8.4 or later. In the interim, restrict deployment of this app to devices with strong device-level security controls, such as managed Android enterprise devices with MDM enforcement, or limit the app's use to non-sensitive workflows. Consider disabling custom URL scheme handling at the device policy level if possible.

Patch guidance

Verify with Genspark's official advisory or release notes whether a patched version addressing CVE-2026-12190 has been released. If available, upgrade Genspark AI Workspace App to the fixed version. For environments where immediate patching is not feasible, implement supplementary controls such as application sandboxing, device-level access restrictions, and monitoring of URL scheme invocations. Test patches in a staging environment before enterprise rollout.

Detection guidance

Monitor device logs for unusual invocations of the Genspark app via custom URI schemes, particularly those containing unexpected parameters or repeated failed authorization attempts. On managed devices, enable endpoint detection and response (EDR) or mobile threat defense (MTD) tools to flag suspicious inter-app communication targeting ai.mainfunc.genspark. Review app permission logs for unexpected data access by the Genspark app. Detection is most effective on devices with comprehensive logging enabled.

Why prioritize this

While the CVSS score is MEDIUM (5.3), prioritization should be moderate-to-low unless the app handles highly sensitive data or is deployed on high-risk device categories. The local-only attack vector and requirement for user-level access significantly reduce real-world risk in centrally managed environments. However, organizations with BYOD policies, shared Android devices, or heavy reliance on Genspark for sensitive business workflows should prioritize patching and compensating controls.

Risk score, explained

The CVSS v3.1 score of 5.3 (MEDIUM) reflects: Attack Vector Local (AV:L) — no remote exploitation; Attack Complexity Low (AC:L) — standard URI manipulation; Privileges Required (PR:L) — attacker must already have user-level access; User Interaction (UI:N) — no user action needed once access is obtained; Scope Unchanged (S:U) — impact is limited to the app and device; Confidentiality, Integrity, and Availability (C:L/I:L/A:L) — partial impact across all three. The score appropriately penalizes the vulnerability for its limited scope while acknowledging the security control bypass.

Frequently asked questions