Active duty, 10+ years of traditional red team, one AI task order under the LLC, OSAI in progress, and 12 months left before I can go all-in. This site documents the crossing in real time — not after the fact.
I'm still active duty. Traditional red team is still my day job — physical and logical, planning and execution, the full ops cycle. The AI work is real but early: one task order through the LLC, OSAI underway, about 12 months before I can shift the balance.
That's exactly why this site exists now, not later. The crossing is more useful to document in progress than in retrospect. The friction points — where traditional tradecraft transfers cleanly, where it misleads you, where AI systems have no equivalent — are sharpest when you're living in both worlds at once.
Most AI security content comes from people who've already crossed. What's missing is the view from the bridge itself. That's what this is. If you're somewhere in the same transition — strong traditional background, starting to take AI systems seriously as targets — take what's useful.
All views are the author's own and do not represent any current or past employer. Content is published in a personal capacity.
Same kill chain. Different substrate. The mental models that built strong network red teamers translate directly to attacking AI systems — once you know which knob maps to which. This is the living index. Each entry expands into its own deep-dive over time.
The standards I actually reference, mapped against each other so the seams show. What each one covers, what it misses, and when to reach for it.
What earns a place in my toolkit and what doesn't. Both worlds, no vendor pitches.
Where to actually break things. Free first, difficulty honest, time-to-flag noted. Self-hosted beats guided every time for building real muscle memory.
Papers, distilled for operators. What matters this week, what mattered last year.
For traditional pentesters going AI, and ML folks learning to think adversarially. Written from the middle of the crossing, not the other side.
Sanitized observations and methodology notes. Technique over name-dropping.
The reflexes that make a strong network red teamer don't transfer cleanly. They transfer well — once you know what to recalibrate. A first attempt at naming the deltas.
Skipping the changelog summary. What I tried, what worked, what's marketing, and where it still has gaps for engagement-grade probing.
Persistence in classic ops is about staying in. In agentic systems, it's about staying influential. Same instinct, different substrate. Worked example with a vector store.
Distilled for people who have to actually exploit or defend this in the next 30 days, not the next conference cycle.
The writing here is personal. The formal work — engagements, task orders, assessments — lives at Crow's Nest, an LLC focused on offensive security across traditional and AI systems. Not actively taking new clients right now, but the work is real.
crows-nest.tech ↗