Roadmap

The RMF Field Guide is a practical guidebook to two NIST frameworks: the Risk Management Framework (RMF) with its SP 800-53 controls, and the AI Risk Management Framework (AI RMF). For each one, it explains what the framework asks of you and ships the documents to act on it, so an organization can learn, build and prove its program. It is written for any organization. Where systems governed by NIST frameworks, such as federal systems, have extra requirements, the guide marks them.

This page shows the plan and where the work stands. For what changed recently, see the changelog; for live progress by control, see guidance coverage.

Two frameworks

The RMF and SP 800-53

The first guidebook covers the RMF from start to finish: the seven steps, every SP 800-53 control, and a program kit to adopt. Every control page will answer six questions, in order, and every answer after the first is something you can copy or follow.

Question	What the guide gives you	Status
Requirement: what does NIST require?	The control text, parameters and assessment objectives, generated from NIST’s official catalog	Available for all 300 controls
Policy: what do we commit to?	A policy clause for the control, assembled into family policies	In progress (Phase 3)
Method: what is the established way to do it?	The NIST publication that shows how, such as the Secure Software Development Framework (SSDF)	Planned (Phase 4)
Procedure: how do we do it here?	Standards, procedures and plans to adopt	In progress (Phase 3)
Implementation: where is it configured?	Technology playbooks for specific platforms	Planned (Phase 6)
Evidence: how do we prove it?	What assessors ask for, and the forms and registers that record it	Available for every control with guidance; growing with each family

Families roll these pieces up into the documents an organization actually adopts: a policy per family (or one consolidated policy), the plans the controls require, and a decision worksheet listing every choice the family forces.

The AI RMF and AI security

The second guidebook does the same for AI: how to apply the AI RMF, and how to adopt, build and run AI securely. It will cover AI in any organization, and AI inside a system authorized under the RMF. It is built only on NIST publications and OWASP’s AI security guidance, with each source’s version and status shown. See Phase 5 for what it will include.

Where it stands

As of October 2026:

• The framework. All seven RMF steps, roles, the ATO package checklist and program variants.
• The controls. All 300 SP 800-53 Rev. 5 controls and their enhancements, with practical guidance for all 177 Moderate-baseline controls and 16 of the 32 program management controls. See guidance coverage for each control.
• The template kit. Version 1.0.0 is released: every template in Word and Markdown, per baseline, with a starter kit for a new program.
• Policies. Family policies for all 20 families: AC, AT, AU, CA, CM, CP, IA, IR, MA, MP, PE, PL, PM, PS, PT, RA, SA, SC, SI and SR. All 287 Moderate-baseline controls have a policy clause.
• Plans, standards, procedures, forms and reports. 46 of the 50 artifacts in the plan, including the System Security Plan, Incident Response Plan, Contingency Plan, Configuration Management Plan, Supply Chain Risk Management Plan, POA&M, Continuous Monitoring Strategy, privacy impact assessment, account management and log review procedures, change request form, component inventory, and the baseline configuration, identification and authentication, audit logging, encryption and patching standards. See all templates.
• Build your program. A staged path from Foundation to Mature, with the artifacts and decisions for each stage.
• The AI RMF guidebook. Not started; it follows the program kit and the methods.

Everything is published as a draft, and marked reviewed only after the author has reviewed it.

What’s next

The work runs in phases. Each phase ends with a review before the next starts, and the phases are ordered by what depends on what, not by date.

Phase	What it adds	Done when	Kit
1 Launch	The public site, generated control pages and quality checks	Done
2 Templates and first kit	The template system, guidance for the priority controls, and the first five family policies with the SSP, IR plan and POA&M	Done, September 2026	v1.0.0
3 Full program kit	The remaining family policies, a consolidated policy, the other plans, standards and forms the controls call for, and practical guidance for every Moderate, program management and privacy control	Every Low and Moderate control has a policy clause, every Moderate, program management and privacy control has guidance, and the program path is complete	v2.0.0
4 Methods and the SSDF	NIST’s how-to guidance shown on each control page, starting with the SSDF, plus secure software development templates	Every family names its NIST methods, and every SSDF practice has a page	v2.1.0
5 AI RMF and AI security	The second guidebook: the AI RMF, securing AI systems through the RMF, NIST and OWASP AI security guidance, and AI templates	The AI RMF pages, the RMF walkthrough, the OWASP pages and the AI templates are published, with a current status table of every source	v2.2.0
6 Industries and technology	Three industry guides, four technology playbooks, and articles	All seven are published and linked from the control pages they cover
7 Depth	Guidance for High-baseline controls, High-baseline clauses, NIST-published crosswalks and automated checks for new NIST releases	Every High control has guidance

Phase 3, the current phase

What remains:

• The SR family policy, including its Privacy-baseline clauses, and the few Privacy-baseline clauses still to add to the AC, AU and IR policies.
• Artifacts for SR: a Supply Chain Risk Management Plan with a supplier questionnaire.
• Artifacts for families already covered: an account management procedure, access request and access review forms and a remote access standard; an identification and authentication standard; an audit logging standard and log review procedure; and an incident handling playbook, incident report form and tabletop exercise kit.
• Guidance for the remaining 52 Moderate controls, and for the program management and privacy controls: how to apply each control, the typical values, the evidence assessors ask for, what is usually inherited and the common findings, family by family, alongside the policies.
• A consolidated policy: one Information Security and Privacy Policy per baseline, as an alternative to 20 family policies. The starter kit will use it.
• The program path: an artifact checklist, and guidance on scaling the program to the organization’s size.
• A template changelog on each template page, then kit v2.0.0.

Methods and the SSDF (Phase 4)

A control says what to do; for many controls, NIST also publishes how it is normally done. This phase adds that how-to layer to the RMF guidebook.

• A methods index: for each control family, the NIST publications that show how to meet its controls, such as SP 800-61 for incident response and SP 800-34 for contingency planning, each with its version and status.
• A Methods section on each control page. A method is linked to a control only where the NIST publication itself cites that control.
• The Secure Software Development Framework (SSDF, SP 800-218) first: a page for each practice in its four groups (Prepare the Organization, Protect the Software, Produce Well-Secured Software, Respond to Vulnerabilities), with the SP 800-53 controls NIST says each task supports, shown on those control pages too.
• Software development templates: a Secure Software Development Policy and SDLC standard organized by SSDF practice, plus a vulnerability disclosure policy, SBOM guidance and an attestation readiness checklist for software producers.

AI RMF and AI security (Phase 5)

A comprehensive guide to applying the AI RMF and to adopting and securing AI, with its own section and entry point.

• The AI RMF itself: an overview and a page for each of its four functions (Govern, Map, Measure, Manage), with their categories and subcategories from NIST’s published framework, each linked to NIST’s AI RMF Playbook.
• The AI RMF alongside the RMF: how the two frameworks fit together, for an organization that runs both.
• Securing an AI system through the RMF: taking a system with AI components (a hosted model, an AI service or an agent) through the seven RMF steps: inventory and categorization, the boundary, what the system security plan records, supply chain, assessment and monitoring.
• NIST’s AI security guidance: the Generative AI Profile, NIST’s taxonomy of attacks on machine learning and their mitigations, and secure development practices for generative AI, which extend the SSDF.
• OWASP’s AI security guidance: the OWASP Top 10 for LLM Applications and for Agentic Applications, explained in the guide’s own words with links to each entry, and the OWASP AI Exchange for application-level threats and defenses.
• AI templates: an AI acceptable use standard, an AI system and use-case inventory, an AI system description appendix for the system security plan, an AI risk and impact assessment, and a third-party AI service review.
• A status table of every NIST and OWASP publication the guide uses, with versions, rechecked at each release. NIST and OWASP both revise their AI guidance often, and the guide follows new editions.

The guide will not invent AI-specific control selections or mappings. NIST is developing SP 800-53 control overlays for AI systems; the guide adds them to the control pages once NIST finalizes each one.

Industries and technology (Phase 6)

How the controls apply under sector rules, and how they are configured and evidenced on specific platforms. The topics below are proposed and may change before the phase starts.

• Industry guides: how the RMF and SP 800-53 apply alongside each sector’s own rules, which controls carry extra weight, and the common findings. Proposed: defense and the defense industrial base (DoD’s RMF instruction, CNSSI 1253, CMMC and SP 800-171), healthcare (the HIPAA Security Rule and HHS 405(d) practices) and financial services (the GLBA Safeguards Rule, FFIEC guidance and PCI DSS).
• Technology playbooks: how specific controls are configured and evidenced on a platform, starting with the controls assessors look at most. Proposed: AWS, Microsoft Azure, Microsoft 365 and Entra ID, and Kubernetes.
• Articles: occasional longer pieces on RMF practice.

Each industry guide and playbook lists the controls it covers, and those control pages link back to it.

What the guide will not do

• Store your data, track your POA&M or act as a system of record. Templates are documents you download and complete in your own tools.
• Copy copyrighted standards or commercial template libraries. Templates are written from the NIST requirement up.
• Give legal or official advice. Confirm requirements with your agency, regulator or program office.
• Run AI tools, or test, evaluate or monitor models. The AI guidebook is guidance and templates only.

Suggest something

The plan changes as the work does. Suggestions and corrections are welcome through the “Edit page” link at the bottom of every page, or as an issue on GitHub.