←
we452,663 posts
we45 avatar

we45

@we45

We are your partners in product security. Creators of @appsecengineer and @orchestronio. AppSec Training | Cloud & Kubernetes | Threat Modeling | DevSecOps

◷ Joined November 2009▧ 533 Photos and videos
we45 avatar
we45@we45·Aug 15

Celebrating the spirit that unites us. Happy 80th Independence Day! 🇮🇳

we45 avatar
we45@we45·May 19

AI agents are already plugged into your systems, querying data, triggering workflows, and making decisions. But they don’t behave like users, and your existing security stack isn’t built to stop them.

Post attachment
we45 avatar
we45@we45·Apr 27

AI agents don’t just assist anymore. They decide. What happens when your AI agents make the wrong decision? You can’t afford to skip this workshop by @abhaybhargav. 🔗 Link in reply

Post attachment
we45 avatar
we45@we45·Apr 24

AI agents are entering your development lifecycle. But your controls haven’t changed. Can your current SDLC handle autonomous developers? If not, this is a must-join. 🔗 Link in reply

Post attachment
we45 avatar
we45@we45·Apr 23

Attackers move daily. Your security assurance should too. o2 plugs directly into your CI/CD pipeline via API — automated scan triggering, programmatic finding management, and continuous validation. Ready to close the gap? Let’s talk.

Post attachment
we45 avatar
we45@we45·Apr 20

You’re not just securing code anymore. You’re securing systems that generate it. Are you securing the output… or the system creating it? You need to join this. 🔗 Link in reply

Post attachment
we45 avatar
we45@we45·Apr 16

You don’t need more tools. You need a new lifecycle. • Validate agent decisions • Add control points • Build visibility into workflows Security has to move with the agents.

we45 avatar
we45@we45·Apr 16

This creates real risks: • Privilege escalation • Unchecked code generation • Autonomous actions with no control points And most teams don’t have visibility into any of it.

we45 avatar
we45@we45·Apr 16

Traditional AppSec expects: ✔ Human reviews ✔ Predictable workflows ✔ Clear ownership Agentic systems break all three. Now you’re dealing with invisible actions and unvalidated decisions.

we45 avatar
we45@we45·Apr 16

AI agents don’t fail loudly. They fail silently. • Wrong prompts → unintended actions • Out-of-context decisions → risky workflows • No audit trail → no accountability You won’t see it coming.

we45 avatar
we45@we45·Apr 16

Your SDLC wasn’t built for AI agents. They don’t just assist anymore. They generate code, make decisions, and act across systems. But your security model still assumes humans. That’s the gap. 🧵