I’m Noah Potti. I mostly work to answer: if a credential, identity, endpoint, infrastructure, or agent is compromised, how far can it reach, and how do you get the teams who own it to close the gap?
Nothing new here
Most of the serious compromises I’ve been part of didn’t need a zero-day. Something already in the environment, a token, a service account, a browser extension, a forgotten permission, had more access than anyone remembered granting it. Much of the work is finding the gap between what a system is supposed to reach and what it can actually reach, and then closing it, especially across teams and at scale, where talking and influencing change is the hard part.
Agents make this harder. The software holding the credential now also reads untrusted text and decides what to do with it. It’s the new browser, except we’re skipping thirty years of hardening and shipping with prompt injection unresolved. An agent needs broad latitude to be useful and still must not be talked into a different task: a read-only assistant that injected text can convince to ship what it found to somewhere it shouldn’t.
The other problem is measurement. Security has little reliable evidence about which controls actually work. In red teaming I’ve watched mature programs, full of controls and checklists, still fall, and most security metrics average bandaids with heart surgery. The papers that convinced me are on the resources page.
Tools
Forwards a site’s WebAuthn request to the YubiKey of the victim. Sign in with a remote key, or test your sign-in and conditional access policies.
Checks leaked credentials to see whether they still work and what they can access. Supports over 100 credential types and uses read-only checks.
Built with the Puck Security folks.
A single executable containing a small LLM and C compiler. Generates, compiles, and runs programs to answer plain-English questions. Runs on Linux, macOS, and Windows.
Scans repositories, packages, and install scripts for signs of malicious code before you run them.
Built at Adversis.
Audits Tailscale ACLs and settings for 57 misconfigurations, with an interactive fix mode.
Built at Adversis.
Writing
- Jan 2027 The Effective Red Team: Adversary Emulation Inside the Enterprise — No Starch Press With Trevin Edgeworth and Jordan Potti. Build a red team that turns successful attacks into better security. Early access ebook is out now; print in January 2027.
- Oct 2026 Ask a Remote Browser to Tap Their YubiKeyTwo browser extensions forward a site’s WebAuthn request to a YubiKey on another computer. No cookies copied. A way to test whether sign-in policy checks the device or just the key.
- Oct 2026 After Red Teaming: What an Adversarial Assurance Function Actually Buys YouRed teams prove assumptions wrong. Adversarial assurance is the harder work after: what had to be true for the path to work, prove the fix mattered, and reopen the question when the company changes.
- Sep 2026 Beyond the Security OrganizationA closed-loop operating model where security evidence changes the right layer of the system and gets revalidated, instead of becoming a ticket. Also as a PDF.
- Jun 2026 Which credential do you rotate first? at Puck SecurityWhy geiger exists: scanners find keys, then leave the question open. Is it live, what does it reach, and in what order do you rotate.
- May 2026 Five findings every AI pen test walks through at AdversisIndirect injection, markdown exfil, tool abuse, cross-tenant retrieval, OAuth scope sprawl. Individually medium; together, critical.
- May 2026 Securing Claude Code for teams at AdversisPermission rules aren’t controls. Managed settings, sandboxes, and the two playbooks that scale from ten engineers to a hundred.
Now
Fractional CISO for a handful of companies through Adversis. AI security assessments for a few more. Building Puck, which is mostly about answering “are we affected?” from the endpoints instead of from a meeting.
Email noah at this domain. Haikus get read first. Everything else gets read eventually.
Elsewhere
GitHub (@thesubtlety, and @noahpotti) · LinkedIn · Adversis · Puck Security