AxiLayer AI | NIST AI 600-1 Generative AI Profile

The framework

The profile federal buyers ask vendors about

What it is

NIST AI 600-1, the Artificial Intelligence Risk Management Framework: Generative AI Profile, published July 2024 as a companion to the AI RMF 1.0. It names twelve risk categories specific to generative AI and more than 200 suggested actions across Govern, Map, Measure and Manage.

Where it is used

Referenced by US federal agency AI guidance, state procurement, insurers and the NIST Cyber AI Profile work of 2025 and 2026. Agencies and their vendors are asked to show which suggested actions they took.

How AxiLayer uses it

Each AI system's monitoring plan lists the risk categories that apply; the scorecard reports the actions as taken or not, with evidence; the passport names the profile.

Mapping

Each requirement, the control that meets it

Every row is a control that is live in AxiSentinel or AxiReady today, with the phase it belongs to.

Requirement	What AxiSentinel and AxiReady do	Phases
Information security (prompt injection, data poisoning, model extraction)	LLM proxy and MCP Gateway in front of every model and tool; injection and poisoning cases in the test suite; behavior guard for drift.	03, 05, 07, 10
Confabulation	Output verification and citation checks in the agentic test suite; human confirmation on high-impact actions.	06, 10
Human-AI configuration	Authority gates, approvals inbox and a stop a person confirms; no AI approves another AI.	05, 06, 08
Data privacy	Per-agent identity and data scope recorded on every action; prompt text never stored, log bodies masked.	04, 09
Value chain and component integration	Discovery of models, connectors and MCP servers; AI-BOM; approval before first use; vendor AI in the inventory.	01, 03
Information integrity	Article 50 style labelling checks and provenance evidence in the suite where content is generated.	10, 11
Harmful bias and homogenization	Fairness and selection-rate tests repeated after every change; findings signed.	10, 11
Dangerous, violent or hateful content; obscene content; CBRN information	Safety cases in the agentic test suite; outbound tools gated; incidents carry a clock.	03, 09, 10
Intellectual property	Component and model provenance in the AI-BOM; retrieval sources in the inventory.	01, 02
Environmental impacts	Cost and usage screen per model and agent; usage statements signed monthly.	09, 11
Govern, Map, Measure, Manage (AI RMF core)	Classify and plan per system; monitoring plan; scorecard; signed passport; re-verification.	02, 09, 11, 12

Sources: NIST AI 600-1 (July 2024); NIST AI RMF 1.0; NIST Cyber AI Profile drafts 2025 to 2026; OMB AI guidance for federal agencies.

Evidence

What the Compliance Passport carries for this framework

Profile mapping per system

Which of the twelve risk categories apply to each AI system and agent, decided at Classify and Plan.

Measured, not asserted

The scorecard shows for each suggested action whether the control fired, held or missed.

Signed passport

NIST AI RMF and the Generative AI Profile listed as frameworks on the passport, re-verifiable free.

Three layers, twelve phases

Twelve phases. One platform. One signature.

Phase 01FindDiscover and score

Phase 02FindClassify and plan

Phase 03ControlApprove every tool

Phase 04ControlName every agent

Phase 05ControlGate every action

Phase 06ControlHold for a person

Phase 07ControlWatch behavior

Phase 08ControlQuarantine and stop

Phase 09ControlMonitor to the article

Phase 10ProveTest and re-test

Phase 11ProveSign the evidence

Phase 12ProveVerify, free, forever

AxiReady finds. AxiSentinel gates, holds, quarantines and stops, Always On or On Demand. A qualified AxiLayer AI auditor signs against NIST AI 600-1.

NIST AI 600-1: the Generative AI Profile

The profile federal buyers ask vendors about

Each requirement, the control that meets it

What the Compliance Passport carries for this framework

See where your agents stand in three minutes