People here keep saying, “We need to rewrite all the human code with AI so it’s secure,” but seem to be leaving out a few steps. A bunch of AI pilled, AI everything startups have been getting popped through web/API bugs this year. There’s clearly another step between “rewrite all the code with AI” and poof, everything is secure.
Companies have maintenance windows where they might patch once a month, or more frequently depending on severity. Many are now patching much more often with less rigorous testing, resulting in some of what we’re seeing now. You move fast and break stuff, but processes will improve over time.
It’s better than not patching and getting wrecked 🤷♂️
I don’t think this is correct. This is exactly like Metasploit and SATAN before it. We have script kiddies again. Every generation is more powerful than the last, but the ultimate goal is automated exploitation.
And just in case that I did not make that clear. What actually happened was that PHP team was unhappy Calif posted a video without details after pre notified Hosters had leaked the information already and everyone was talking about it already
A man and his Muse agent are suing Meta. This is a really wild story about 1) how one man has handed over his entire backend business to Meta's AI 2) how courts are going to be overwhelmed by randoms and AI 3) Meta's Muse coming back to haunt it 404media.co/a-man-and-his-mu…
GreyNoise is tracking #java4shell (CVE-2026-114938) with a live dashboard here: java4shell.events.greynoise.…
At this time, we are not aware of successful exploitation in the wild. We believe this will be challenging to mass exploit at least initially due to the need for the exploit to be tailored for each targeted system build.
Global Secure Access and Edge Management Service had a baby 🎉
Welcome to Microsoft's Secure Enterprise Browser :p
Still a few caveats, but you may have scenarios that no longer need the GSA client to access private web apps or secure Internet taffic
learn.microsoft.com/en-us/en…
1/ Java4Shell: Great to see @calif_io write-up discussing the hard parts, not just showing a working PoC.
I read it carefully, trying to learn all the tricks.
Memory leaks can defeat ASLR, but the leak requires careful heap grooming first.
Elastic Security Labs Threat Command detected a credential-stealing supply-chain attack using a malicious version of the Microsoft project agent-framework (1.26.1 and 1.26.2) at 3:41 PM EST on November 17, 2026. There appears to be an additional file (_telemetry.py and af_runtime_diag.py) added to the packages, aimed at Azure cloud workloads and CI/CD build agents.
It’s disguised as a harmless “runtime diagnostics helper”, but steals environment variables, credential files, OIDC tokens, and managed identity tokens. The attack chain exfiltrates to the endpoint: webhook[.]site/d381fd80-ec62-42ad-8f5c-d1b15103c9bb.
If you installed agent-framework (1.26.1, 1.26.2), we recommend removing it, pinning to 1.26.0, and rotating any credentials accessible from that environment, since the release steals secrets and Azure tokens.
Package URL:
pypi.org/project/agent-frame…pypi.org/project/agent-frame…
User-Agent: af-runtime-diagnostics/1
Heads up! The official @Microsoft Python package "agent-framework" that's hosted on @pypi has been compromised. Versions 1.26.1 and 1.26.2 are both malicious.
opensourcemalware.com/blog/a…