CVE-2026-60094: Vinchin Backup & Recovery Heap Buffer Overflow (Remote Denial of Service)

Technical summary

The vulnerability is a classic heap buffer overflow (CWE-787) in Vinchin's agentlink_server service. The service accepts TCP connections and reads packet body data using an unchecked body_len field from the packet header. This length value is passed directly to a recv() call without validation, allowing an attacker to specify a length much larger than the allocated heap buffer. When the service attempts to fill this oversized buffer, it overwrites adjacent heap memory. The overflow magnitude can reach approximately 4 GiB if an attacker sets body_len to its maximum value, though practical exploitation depends on heap layout, memory availability, and the service's internal buffer management. The unauthenticated, network-accessible nature of the agentlink_server makes this remotely exploitable with no prior access required.

Business impact

A successful exploit disables the Vinchin backup agent on affected systems, preventing backups from completing and potentially rendering recovery unavailable during a data loss incident. Attackers can cause repeated crashes to maintain the denial of service. While immediate code execution is not guaranteed, memory corruption could enable privilege escalation or lateral movement within the backup infrastructure. Organizations relying on Vinchin for backup and disaster recovery face reduced resilience. The impact is most severe in environments where the backup agent runs on production systems or critical infrastructure.

Affected systems

Vinchin Backup & Recovery versions up to and including 9.0.0.86562 are affected. The vulnerability exists in the agentlink_server service, which listens on the network. Any deployment of these versions with the agentlink_server exposed (whether directly or via internal network) is vulnerable. The vendors_products field in the source data is empty, so consult Vinchin's official documentation or advisory for precise affected product lines, editions, and component deployments.

Exploitability

Exploitability is high. The attack requires only network access to the agentlink_server port—no authentication, no user interaction, and no special privileges. An attacker can craft a malicious packet from any network position that reaches the service. The flaw is straightforward: unchecked length validation followed by a direct recv() call. Exploit development is feasible; however, the impact (crash or memory corruption) is more predictable than achieving reliable code execution. The vulnerability is not currently listed on CISA's Known Exploited Vulnerabilities catalog, but the simplicity of the bug and its unauthenticated nature make it attractive to threat actors.

Remediation

Immediately identify all systems running Vinchin Backup & Recovery through version 9.0.0.86562. Check Vinchin's official security advisory and product portal for a patched version addressing CVE-2026-60094—verify the patch version against the vendor's guidance, as the source data does not specify a fixed version number. If a patch is available, prioritize deployment to agents protecting critical data. If no patch is yet available, isolate agentlink_server from untrusted networks, restrict TCP access to the service port via firewall rules, and monitor for unusual connection patterns or agent crashes. Consider disabling the service temporarily if an alternative backup method exists and the risk is deemed unacceptable.

Patch guidance

Contact Vinchin directly or consult their security advisory page for the availability and version number of the patch for CVE-2026-60094. Once a patched version is confirmed, test it in a non-production environment to validate backup and recovery functionality before rolling out to production agents. Given the network-exposed nature of the service, prioritize patching agents that are internet-facing or accessible from untrusted segments. Implement a phased rollout to minimize service interruption if the patch introduces any compatibility issues.

Detection guidance

Monitor agentlink_server processes for unexpected crashes or restart loops, which may indicate exploitation attempts. Enable network intrusion detection on inbound traffic to the agentlink_server port; look for malformed TCP packets with unusually large body_len fields or packets that deviate from the expected protocol structure. Inspect the backup agent's logs for errors related to memory allocation or buffer handling. If available, deploy endpoint detection and response (EDR) tools to capture heap corruption events or unusual memory access patterns on systems running the agent. Baseline normal traffic to the service and alert on anomalies.

Why prioritize this

Although listed as MEDIUM severity (CVSS 6.5), this vulnerability warrants high prioritization due to the combination of unauthenticated network access, lack of complexity, and the critical nature of backup infrastructure. A successful attack disables a core resilience function. In environments where backup agents are exposed to less-trusted networks or the internet, prioritization should be elevated. The absence of a known public exploit does not diminish risk; the vulnerability is straightforward enough that adversaries will develop tooling quickly if they identify vulnerable deployments.

Risk score, explained

The CVSS 3.1 score of 6.5 reflects a MEDIUM base risk: network-accessible (AV:N), no authentication required (PR:N), no user interaction (UI:N), affecting only the target service (S:U), with limited integrity impact (I:L) and availability impact (A:L). The score does not account for the criticality of backup infrastructure or the potential for heap corruption to cascade into further compromise. Organizations should consider a higher contextual risk level if this service is exposed to adversary-controlled networks or if the affected backup system protects high-value data.

Frequently asked questions