Loki Intelligence — Security Briefs · Published
Daily Security Brief: Xen Orchestra, Railway, AI Agent Risk
Today's reports center on exploited access paths: an authentication bypass in virtualization management, stolen platform secrets, and compromised cloud and AI inference infrastructure. Teams should prioritize patching exploited systems, rotating and scoping credentials, and reviewing the identities and permissions granted to automated agents.
Signal 01 · The Hacker News
Hackers Exploit Critical Xen Orchestra Auth Bypass Flaw to Gain Admin Access
Vates VMS warned that a critical authentication bypass in Xen Orchestra is being exploited in the wild. The flaw did not yet have a CVE identifier at the time of the report.
Why it matters: Restrict network exposure of management consoles and apply vendor guidance promptly, since administrative access to virtualization tooling affects every workload it manages.
Source: The Hacker News
Signal 02 · BleepingComputer
Railway breach leads to $740,000 in fraudulent AI charges
Railway said attackers exploited vulnerabilities in its infrastructure to steal customer authentication secrets, resulting in roughly $740,000 of fraudulent AI usage. Credentials for up to 100,000 workspaces may have been exposed.
Why it matters: Rotate secrets stored on third-party platforms after such incidents and set spending limits and usage alerts on AI provider keys.
Source: BleepingComputer
Signal 03 · The Hacker News
We Built An Agent That Broke All of Our Security Rules. Here's What We Learned.
A vendor describes building an internal AI agent that ended up violating its own guidance on agent identity, scoped access and separate traceability.
Why it matters: Give each agent its own identity, least-privilege access and audit trail distinct from any human user, even for internal tools.
Source: The Hacker News
Signal 04 · CISA
CISA Adds Six Known Exploited Vulnerabilities to Catalog
CISA added six vulnerabilities to its Known Exploited Vulnerabilities Catalog based on evidence of active exploitation. Listed entries include a Cisco BroadWorks XML external entity issue and a Barracuda Networks OS command injection issue.
Why it matters: Check your asset inventory against the new KEV entries and prioritize remediation of any affected products.
Source: CISA
Signal 05 · BleepingComputer
Hackers exploit trusted AWS access to breach multiple cloud environments
A threat actor used compromised IAM credentials tied to a Qubole big data platform's AWS account to access and exfiltrate data from several customers' AWS environments.
Why it matters: Review cross-account trust relationships and third-party IAM roles, and limit their permissions and monitor their activity.
Source: BleepingComputer
Signal 06 · The Hacker News
SGLang RCE Exploit Breaches NEAR AI, Steals Data From 18 Agents and About 100 Users
An attacker exploited a remote code execution flaw in the open-source SGLang inference engine and ran code on a NEAR AI inference server. Requests processed on that server that day could also have been visible to the attacker.
Why it matters: Keep inference engines patched, isolate them from sensitive networks, and treat prompts and agent data passing through them as exposed if a host is compromised.
Source: The Hacker News
Signal 07 · BleepingComputer
Critical Xen Orchestra auth bypass flaw exploited to hijack VM servers
BleepingComputer reports the critical Xen Orchestra authentication bypass is being exploited to obtain administrator privileges and potentially take over XCP-ng virtualization hosts.
Why it matters: Audit Xen Orchestra instances for unexpected administrator accounts or activity and limit access to trusted networks until patched.
Source: BleepingComputer
Brief sources
Related briefs
- What Is AI Agent Security Testing?
- AI Red Teaming vs Traditional Pentesting
- MCP Security Testing Checklist
Relevant Loki service: Orvyn — AI agent security private preview. Planning a test? See what a pentest costs and how a scanner, a freelancer, a pentest firm and SKYEN compare.