FreeScout CVE-2026-41906

Requires valid low-privilege agent authentication credentials (PR:L in CVSS vector) for any mailbox within the FreeScout instance. The target FreeScout installation must use multi-mailbox configuration with role-based access controls that restrict agents to specific mailboxes (single-mailbox deployments lack cross-tenant attack surface). The attacker must have visibility to at least one conversation within their authorized mailbox to obtain a valid conversation_id for manipulation. Exploitation is limited to reassigning visible conversations to customers in other mailboxes; the agent cannot create new conversations or directly access the unauthorized mailbox's conversation list (scope remains unchanged, S:U). The backend must be running a version prior to 1.8.214 lacking server-side customer access validation in the conversation_change_customer controller action.

Real-world risk is MODERATE despite CVSS 7.1 (High). The CVSS vector (AV:N/AC:L/PR:L/UI:N) indicates network-reachable exploitation with low complexity, but requires valid agent credentials (PR:L) and only affects confidentiality partially (C:L) while maintaining high integrity impact (I:H) from unauthorized data modification. The integrity impact reflects ability to manipulate conversation-customer associations, potentially exposing sensitive customer data across tenant boundaries or enabling social engineering attacks. Exploitation is constrained to authenticated insiders or compromised agent accounts, reducing the attack surface compared to unauthenticated RCE. No EPSS score provided in dataset. Not listed in CISA KEV, indicating no confirmed active exploitation. The SSVC framework would classify this as 'Scheduled' priority: exploitation requires authentication (not automatable), technical impact is partial (data manipulation within help desk scope), and multi-tenant boundary violations have clear business impact for SaaS deployments. The bundling of five related authorization fixes in a single release suggests vendors discovered these through internal audit rather than active exploitation. Priority escalates for multi-tenant FreeScout deployments where mailbox isolation is security-critical; single-tenant or trusted-agent environments face lower practical risk.

An authenticated agent with legitimate access to Mailbox A (Sales) intercepts the Change Customer dialog's network traffic and identifies the conversation_change_customer API endpoint. The agent crafts a POST request with a conversation_id visible in their mailbox but substitutes the customer_email parameter with a high-value customer from Mailbox B (Executive Support), which they cannot normally access. The backend processes the reassignment without verifying mailbox boundaries, binding the conversation to the unauthorized customer. The attacker can now view the conversation's full history including sensitive customer details from Mailbox B, exfiltrate confidential support communications, or manipulate conversation data across tenant boundaries. Exploitation requires only standard HTTP client tools (cURL, Burp Suite) and valid agent credentials.