NullGhost_0x

Cybersecurity engineer. Building NullGhost — offensive security & red team. Sharing a security tool worth knowing, daily.

[NullGhost_0x @Nullghost02014]

@ThrillScience Thanks for the link! That explains the sampling side. What are you using to stream and process the raw I/Q data in your setup?

[NullGhost_0x @Nullghost02014]

@poteto Are credentials kept isolated from the model’s context? I’d be interested to hear how that flow is protected against prompt injection on pages the agent visits.

[NullGhost_0x @Nullghost02014]

Vulnerabilities keep surfacing. AI-driven security assessments are dramatically shortening the path from discovery to remediation.

[The Hacker News @TheHackersNews]

‼️ FOUR critical GitLab RCE flaws traced back to just THREE bugs in Ruby's regex engine.

Ruby has now fixed them, but this isn't just a GitLab issue.

Other Ruby apps that compile attacker-controlled regex patterns are also affected.

Details ➜ https://t.co/E4NxZyVZcq

[The Hacker News @TheHackersNews]

‼️ FOUR critical GitLab RCE flaws traced back to just THREE bugs in Ruby's regex engine.

Ruby has now fixed them, but this isn't just a GitLab issue.

Other Ruby apps that compile attacker-controlled regex patterns are also affected.

Details ➜ https://t.co/E4NxZyVZcq

[NullGhost_0x @Nullghost02014]

@Huntio Useful distinction between exposed PHP versions and confirmed exploitability. What evidence are you using to separate exploit staging from successful execution?

[NullGhost_0x @Nullghost02014]

@three_cube Curious how it handles authenticated sessions and business logic testing. Does it validate findings with reproducible evidence, or leave that step to the operator?

[NullGhost_0x @Nullghost02014]

@DarkWebInformer Vulnerabilities are just popping up all over the place, in the most ridiculous way.

[broke boy @0xfuckpoverty]

This is f*king insane. SpaceXAI engineer Lauren Tan just opened up the setup that turned 2 years of rejections into a $500K offer

the story: 250 applications, zero offers. then he showed up running 20 agents.

in this 1-hour workshop, Lauren walks through how to build a Grok Bot team that delegates the work and checks its own results.

Grok Bot → Chief of Staff → Managers → Workers

a chat window gives you one bot. an agent team gives you a system.

this free workshop is worth more than most $800 agent engineering courses.

bookmark it and watch it today.

Then read the full article below ↓

[NullGhost_0x @Nullghost02014]

@bcs_erictaylor Just yesterday I was showing that it's easier than going to pick up your delivery

[NullGhost_0x @Nullghost02014]

@0xnicc0 This is very crazy opus 5.5

[NullGhost_0x @Nullghost02014]

@shirosignal The most awesome Lauren always does key things

[NullGhost_0x @Nullghost02014]

@three_cube Your screen can be air-gapped. Its electromagnetic emissions aren’t. TEMPEST attacks are still wild.

[NullGhost_0x @Nullghost02014]

@markgurman It's better to wait and see so it doesn't turn out to be a huge disappointment; honestly, I have a lot of faith in this device.

[NullGhost_0x @Nullghost02014]

@SecurityTrybe the final transformation

[NullGhost_0x @Nullghost02014]

@fr0gger_ We are entering a new era of AI-driven work.

[NullGhost_0x @Nullghost02014]

@Dinosn https://t.co/LArIjkSq60

[NullGhost_0x @Nullghost02014]

https://t.co/ReDqjzxmmm

[NullGhost_0x @Nullghost02014]

AgentDojo: a dynamic environment to evaluate attacks and defenses for LLM agents.

#CyberSecurity #AISecurity #LLMSecurity

[NullGhost_0x @Nullghost02014]

https://t.co/D0MZEFfOl2

[NullGhost_0x @Nullghost02014]

The open source security engine for AI agent and supply-chain trust.

#CyberSecurity #AISecurity #AgentSecurity