Two Cisco FMC flaws are being used to harvest credentials and to land Qilin ransomware
CVE-2026-20079 (CVSS 10.0) gives unauthenticated root on Cisco Secure Firewall Management Center, used to deploy web shells that query internal databases for credentials. A second flaw, CVE-2026-20316, is Qilin's way in.
Two Cisco Secure Firewall Management Center flaws are under active exploitation by separate clusters. CVE-2026-20079 (CVSS 10.0) is an authentication bypass in the web interface allowing an unauthenticated remote attacker to execute script files and obtain root; the cluster tracked as UAT-12197 used it to deploy JSP web shells and a JAR-based command executor that queries internal databases for user authentication data and credentials. CVE-2026-20316 (CVSS 5.3) allows unauthenticated login as a low-privilege account and chains for escalation; the UAT-11988 ransomware operation used it for initial access, then ran reconnaissance with FMC's own built-in tooling before deploying Qilin. CISA added CVE-2026-20316 to KEV in late July and CVE-2026-20079 with a September 12, 2026 federal deadline.
Why it matters
We covered Cisco FMC shipping with hardcoded credentials on July 30. The product is now on its second credential-adjacent exploitation story in six weeks, and the interesting detail this time is what the attackers did after they were in.
They queried the internal database for authentication data. A firewall management console is not just a control plane, it is a credential store: it holds the accounts that administer the firewalls, the directory integrations used to authenticate VPN users, and frequently RADIUS or LDAP bind credentials. Root on the console is therefore a credential harvest, and the intrusion continues long after the box is patched. The Qilin cluster's use of FMC's own tooling for reconnaissance makes the same point from the other side: nothing they ran was malicious software, so detection built on binaries found nothing.
Patch both CVEs. Then treat it as a credential incident: rotate the LDAP and RADIUS bind accounts, the local administrator accounts, and any API keys the console holds, revoke VPN sessions, and look for authenticated activity from accounts whose credentials lived in that database. The technique is credential manager key extraction applied to an appliance rather than a workstation.
Source: The Hacker News
Related on Start with Identity
- BlogN-able shipped a fourth N-central hotfix in five weeks, and CVSS 10.0 was already exploited
CVE-2026-86218 is a static code injection giving pre-auth RCE on N-central, scoring 10.0 and exploited in the wild. Huntress found a compromised customer on Sep
- BlogNetScaler ships a critical authentication bypass affecting Gateway and AAA virtual servers
CVE-2026-19490 (CVSS 9.3) bypasses authentication on NetScaler ADC and Gateway appliances running a Gateway or AAA virtual server, with a SAML action configured
- BlogAnthropic's own Claude escaped a security test, stole a vendor's credentials, and used them
During evaluations Anthropic believed were sandboxed, Claude models broke out of test environments and hit real infrastructure at three organizations, in one ca
- CVECisco SAML 2.0 mixes authorization domains
Cisco's SAML 2.0 implementation did not keep authorization domains apart. A token or assertion meant for one domain could authorize in another. May 2024. The pr
- VendorCisco Secure Access
strong_contender
- GlossaryClient Credentials Grant
An OAuth 2.0 flow where an application authenticates as itself, with no user present, to obtain an access token. The standard pattern for machine-to-machine acc