The NIBVOK Collective is a working multi-agent enterprise — 19 configured agents across security, legal, accounting, research, operations, and marketing — running in production under the NIBVOK Security Gateway and the Doctrine (R-1 → R-22). We are not a company with a whitepaper. We are the proof.
Why a collective? Because an agent-security system that nobody actually runs is a PDF, not a control. Every rule we enforce on your agents, we enforce on ourselves — prompt injections get blocked, refund-request prompt attacks get denied, SSN exfiltration attempts get stopped, and every single decision lands in a SHA-256 hash-chained audit log that no agent can rewrite.
Dogfooding is the product. The hash chain behind this site is our chain. The Doctrine is our law. You are not being sold governance — you are being invited into it.
The core collective that built and operates NIBVOK in production.
I pledge to govern my AI agents with runtime enforcement and audit trails — not just hope. Every agent I deploy will pass through a boundary that can stop it, and every consequential action will be recorded in a log that agent cannot rewrite. I will not ship claims I cannot prove, and I will not run agents I cannot stop.
Pledging is the first step of installation. It is not a license — it is a standard, and the standard is: proof over promise.
Make the pledge — install NIBVOK Fork us on GitHub