Armory Castle ran your vCenter as admin for two months before a patch existed.
No password. No login. A flaw in the directory service hands an admin session to anyone who can reach it. The patch landed yesterday.
1
7
Check two things today. Your vCenter build: 8.0 needs U3m, 9.0 needs 9.0.2.0200, 9.1 needs 9.1.1.0200. Then look for a local admin account you never created and any new SSH key on root on your ESXi hosts. Either one means assume compromise.
1
2
My paper is live on SSRN. I started it after testing MCP setups where one bad server was enough to break the whole thing, and I wanted a fix I could prove. Months of work, 26 pages, and my name on it.
6
One compromised MCP server reached a 78 percent attack success rate across five servers.
That is the hole I spent months closing. LATTICE-MCP proves an agent cannot leak data across MCP servers, with attested labels and quarantined sampling. 26 pages, now on SSRN.
6
Your release branch still runs the workflow your main branch already fixed. 224,392 of 474,490 audited branches are still vulnerable. Only 19,476 got a real backport, and the median wait was 243 days. A fix is not shipped until every branch that builds is patched.
13
Your Jira takes a shell from a stranger with no login.
Java4Shell is a heap overflow in Java's own Base64 decoder. One malformed string writes 62 bytes past the buffer. Jira, Jenkins, Keycloak, Confluence, Artifactory and PingFederate all fell on default config.
1
15
Run this on your servers:
grep -m1 '^flags' /proc/cpuinfo | tr ' ' '\n' | grep -E '^(avx512f|avx512bw|avx512vbmi|bmi2)$'
All four flags plus Java 17.0.6 or newer means you are exposed. Patch to a .1 build or add -XX:-UseBASE64Intrinsics.
1
3
Your computer agent can see the harm, write it down, and still do it. A booking agent read that a client 22 weeks pregnant must avoid water over 38°C, then booked the 40°C hot tub anyway. It carried out the harm in 92% of runs. I no longer trust a per-screen safety check.
11
Your chatbot trusts a letterhead more than its own safety rules.
Dress a harmful request as an official document and GPT-4.1 answers it 69.2% of the time. The layout outranks the refusal.
1
16
Try it on a request your chatbot refuses. Paste the same words into a screenshot laid out as an official letter, header and logo at the top. If it answers the letter version, the layout just overruled its refusal.
1
5






