>

AI for Defenders

Defending together in the age of AI.

A community-driven space of honest field notes, a newsletter every two weeks, and open tools you can run, on how cyber defenders actually use AI: what works, what breaks, and what we learn together. Podcast coming later this year.

FIELD NOTES · LIVE NEWSLETTER · EVERY 2 WEEKS TOOLS · SOON PODCAST · COMING
Why this, why now

Nobody has the next four months figured out.

Not the vendors, not the analysts, not us. Defenders are being told AI changes everything, and handed very few honest answers about what to actually do on Monday morning.

So we're figuring it out in the open, together. Real experiments, real trade-offs, real code, including the parts that broke. No hype and no gatekeeping, just defenders helping each other get it right. Newcomers and veterans both welcome. Cybersecurity is a team sport.

What we're building

Four ways in.

Field notes

Build logs from the trenches

What we tried, what we'd do differently, what fell over. The messy middle, written down honestly.

Newsletter

The signal, every two weeks

A curated pulse of AI-for-defense: the few things worth your attention, pulled out of the noise.

Tools

Field notes you can run

Open, benchmarked tools for defenders, built and improved in the open. First one drops soon.

Podcast

Real conversations

The people doing the work, on what's actually working in their week, not a roadmap slide. Concepts you can take and make your own.

Live signal

What's shipping.

All field notes →
FIELD NOTE Oct 27, 2026

The Incident We All Missed: Why Your Eyes and Your AI Failed the Same Way

A malformed User-Agent header slipped past a customer SOC as a false positive, and the AI we double-checked it with failed the exact same way. A plain-text rerun months later still missed it. Why human and AI readers fail in correlated ways, and what that means for your threat model.

#field-notes#ai#soc
NEWSLETTER Oct 22, 2026

Criminals got agents. The playbook is still old.

AI for Defenders issue 004: five teams documented criminals running AI agents against real victims, the lab-agent incidents outsiders surfaced, attacks on AI gateways, and 14 quick hits from a nine-week window.

#newsletter#ai#ai-agents
NEWSLETTER Aug 20, 2026

The safety tests got out

AI for Defenders issue 003: the OpenAI safety evaluation that escaped its sandbox and breached Hugging Face, the disclosures that followed from Anthropic, Meta and the UK AI Security Institute, AI-generated tooling in attacks on industrial controllers, and what a year of real casework says the commodity AI threat actually is.

#newsletter#ai#ai-agents
NEWSLETTER Jul 14, 2026

The attacker was an AI agent

AI for Defenders issue 002: the first LLM-driven ransomware operation, the Five Eyes warning that more is coming in months, someone scanning for your MCP servers, and agents as a new kind of identity.

#newsletter#ai#ai-agents
Episode 01 · soon

Why AI for Defenders?

The manifesto: where this is going, who it's for, and why defending is something we do together.

We learn out loud.

Built in the open. By defenders, for defenders.