As reported by SecurityAffairs, the MikroTrick attack chain demonstrates how two RouterOS vulnerabilities — CVE-2026-86060 and CVE-2026-67279 — can be combined to bypass SSH authentication entirely and grant an attacker full administrative access to MikroTik devices without credentials. MikroTik shipped patches on September 3, 2026, but withheld technical details, a decision that proved futile when CERT Polska researchers diffed the patch packages and reconstructed the exploit chain within hours.
Vulnerability Summary
| CVE | Component | Impact | Patch Status |
|---|---|---|---|
| CVE-2026-86060 | RouterOS SSH authentication | Authentication bypass via rekey manipulation | Patched Sept 3, 2026 |
| CVE-2026-67279 | RouterOS user management | Privileged account creation without auth | Patched Sept 3, 2026 |
Severity: Critical (chained). No explicit CVSS scores were published in the source, but the combined impact equals full unauthenticated administrative compromise.
Affected Products: MikroTik RouterOS versions prior to the September 3, 2026 security update. Specific version ranges were not enumerated in the source reporting.
Active Exploitation: Yes. Forum logs showed attackers using a username beginning with a hyphen (-2) followed by automatic creation of a privileged ops account — a pattern MikroTik itself appeared to know about, given the patch's built-in detection and disabling of rogue ops accounts.
Why This Matters
The core technical lesson is deceptively simple: SSH's layered authentication model is only as trustworthy as the implementation that enforces it. RouterOS handled SSH rekey events — routine mid-session encryption renegotiation — incorrectly when they occurred during the authentication phase. Complete the rekey before auth finishes, and the server jumped ahead to channel handling as though authentication had already succeeded. That logic error is the kind of state-machine flaw that automated fuzzing and protocol conformance testing should catch but manual review frequently misses.
The second vulnerability then allowed creating a fully privileged account through that unauthenticated session. The chaining is straightforward but devastating: no password, no SSH key, no valid credentials of any kind — just full admin.
MikroTik's decision to ship a patch labeled merely an "important security update" without describing what it fixed represents a failed experiment in silent patching. The gap between patch availability and public technical detail is now measured in hours, not weeks, when motivated researchers have AI-assisted diffing tools and public forum telemetry to work with.
Who Is at Risk
Shield53 Recommendations
Immediate Actions
- Patch all RouterOS devices immediately to the September 3, 2026 firmware release or later. Treat this as emergency maintenance, not a routine update cycle.
- Audit for indicators of compromise: Check for unauthorized accounts, particularly one named
opsor any username beginning with a hyphen or space. Review authentication logs for failed login attempts that correlate with subsequent account creation events. - Restrict SSH access: Disable SSH on WAN interfaces. Limit management access to a dedicated VLAN or VPN. If SSH must remain enabled, restrict it to specific source IPs via firewall rules.
- Disable unused services: Reduce attack surface by turning off Winbox, FTP, Telnet, and any API services that are not actively required.
- Review device configurations: Look for unexpected NAT rules, firewall exceptions, or VPN tunnels that an attacker may have added for persistence after gaining admin access.
Strategic Actions
- Inventory all MikroTik assets across your environment. You cannot patch what you have not catalogued. Edge devices are frequently deployed and forgotten.
- Establish a firmware monitoring process for all network infrastructure vendors. Subscribe to MikroTik's notification channels and track release notes even when they are vague.
- Apply network segmentation so that compromised edge devices cannot pivot freely into internal segments.
The broader industry takeaway: silent patching as a vulnerability disclosure strategy is effectively dead. AI-assisted binary diffing, public log sharing, and collaborative research communities mean that the technical details behind any patch will be reconstructed rapidly. Vendors who ship fixes without transparent advisories are not protecting their customers — they are simply delaying informed defensive action by a few days while attackers who reverse-engineer patches get a head start.