Brief
Jacob and Daniel recap CS5, where attendees were still learning that the CMMC pause did not suspend their DFARS 7012 and 800-171 obligations. They warn of future False Claims Act settlements covering the pause period, discuss how solicitation bonus points could act as a requirement, and argue against less-detailed, CSF-style verification, citing a CISA advisory and MITRE ATT&CK mappings. Audience Q&A covers CUI flow-down, DIBCAC’s most common findings, assessment automation, FAR CUI scoping, SPRS verification, and FedRAMP High vs. ITAR.
Key Takeaways
- The pause did not suspend the requirements. DFARS 7012 and 800-171 still apply, and some CS5 attendees were only now realizing it.
- Expect False Claims Act exposure. Jacob predicts future settlements will include the pause period, with no “CMMC was suspended” defense.
- Bonus points can work like a requirement. GAO found the suspension didn’t bar Army MAPS bonus points, which could be an interim way to reward certified companies.
- Primes turn start dates into deadlines. DoD never set a universal November deadline, but primes did, so any new date will likely become one too.
- Detail beats summary. The CSF and Brilliant at the Basics lack verification criteria, so the hosts argue requirements should get more specific, not less.
- Ask about CUI carefully. Asking “is this CUI?” tends to expand the marking, so confirm your own flow-down determination instead.
- Know DIBCAC’s top findings. FIPS-validated cryptography, MFA, vulnerability scanning, audit logging, and baseline configuration are common “other than satisfied” items.
- Verify your MSP. Primes and assessors should confirm a provider actually performs its duties and can produce a shared or customer responsibility matrix.
- FedRAMP High is not ITAR compliance. It certifies baselines, not US-person access or data location.
Transcript
Welcome
Jacob: Hey, everybody, it is Friday. It is hotline time. We’re live on LinkedIn. We’re live on YouTube. You can find us at cuihotline.org. You can fill out the form, you can call the number and leave a message, you can send us a DM, or you can find us at summit7.us. If you’re at CS5 right now, we’re not there, but you can find the rest of the Summit 7 team hanging out. Daniel and I were there yesterday, so we saw some people, which was always a good time. Gave some presentations, which is always a good time. Talked to some contractors, which is always the best time. And yeah, made it back from DC despite the rain. It’s raining in LA, if you guys didn’t know, so it’s always a crisis when that happens. But yeah, whatever.
The team is still at CS5, so if you’re at CS5 watching this, go say hello. We have a bunch of people hanging out there today. There you go. Daniel, weekend review: what was the recap of CS5?
CS5 recap: the pause didn’t end the requirements
Daniel: Yeah. It was interesting. When you go to events like this, and this is honestly the only real CMMC-related event that I’ve seen during the pause, outside of, I think, CMMC Midwest or something of that nature in Oklahoma, you always wonder who you’re going to run into. A ton of ecosystem people, right? People we’ve known for years and years and years. I got to hear my good friend Ryan Bonner give a great talk, him and Michelle.
But then you hear some hallway chatter, and this is the best bit, right? Just mosying by, what are people thinking and saying? I ran into some people that had really no idea what CMMC was but had just immediately been tasked to go fix the problem. Then I ran into somebody. There was a conversation overheard at one of the tables where someone was on the phone with what appeared to be a leader of some sort in their organization, and they were trying to get clarity on, “Wait, so CMMC’s paused? I didn’t think that meant we had to do anything.” This was an attendee at the conference, learning in real time, as far as I could tell: “Oh no, the pause doesn’t mean CMMC’s dead, and I actually have to do something.”
Which kind of dovetails into another conversation I had directly with somebody. They were talking about how certain contracts are basically looking to make CMMC a requirement before they will disclose the RFI or RFP to them. Not as a requirement of contract award, which the CIO has suspended, but as an element of visibility into what the RFP might even be. So, pre-RFP release, validating you are who you say you are, because the data package they’re sending you is, in their eyes, CUI, and they want to make sure you’re compliant to receive it.
And then a whole other conversation, and I’m just rattling these off because there were so many interesting conversations. A contracting officer said, “Yeah, a lot of people don’t know there’s two tabs in SPRS. One has your old score, if you will, and one has your CMMC Level 2 self-assessment. So I had a list of proposals I was looking to review. Before I laid eyes on any of them, I just went into the second tab to see if they had an entry. There was only one that had the entry, and I awarded the work to them because they were the only one that met the requirements that I had in the RFP.”
So yeah, it’s just fascinating little tidbits that you hear left, right, and center. We were honestly left grasping at straws a little bit. Matt Travis did as good a job as he could, given the keynote on the first day. The CIO wasn’t present. So the general direction of CMMC is still where it was last week, through the hotline and through the conversations we’ve been having. But the ecosystem is looking for Rev 3 of 171. FAR CUI even talked about things like NIST CSF as part of that, as well as a lot of AI conversations, and that’s just kind of a global thing right now, right? Governance around AI, no matter what industry you’re in. So there were a lot of interesting conversations, to say the least. But my favorite one is “the pause didn’t kill CMMC.” They didn’t realize in real time at the conference that, yeah, they actually still have to do something related to that.
False Claims Act risk during the pause
Jacob: So yeah, in the not-so-distant future, when the pendulum swings back in the other direction, with the government getting back to focusing on the things it’s supposed to be doing, there will be, mark my words, DOJ False Claims Act settlements regarding cybersecurity noncompliance during the period that includes the CMMC suspension. People will have been attesting to DFARS 252.204-7012. They will not have been doing it because they thought that CMMC’s suspension had anything to do with those requirements, and then they’re going to end up writing a check to the government and a whistleblower for hundreds of thousands or millions of dollars. I guarantee you that settlements in the future will include this time period. They will not say, “Well, this time period doesn’t count because CMMC was suspended.” It’s not how it works. It might take a while, it might not take a while, but eventually we will see those settlements. So take your risks accordingly, everybody.
No DoD presence at CS5, and the messaging gap
Jacob: And like we were talking about before the show, as far as CS5 goes, nobody from the CIO’s office showed up. There were zero representatives from the DoD at the entire conference. So it’s just another iteration. I’ve said this every day for the last, oh, geez, hold on, let me put it back up here, every day for the last, is it 102, 103, every day for the last hundred-some-odd days: just because you have a new leadership team in there has not fixed the culture of “they don’t think they need to repeat themselves if they don’t have anything new to say.”
The number one thing you have to do is keep telling people that the requirements are still in effect. You can’t just say it once. You have to say it over and over and over again. If you’d come out and hung out with everybody, you would know that if you don’t tell people constantly that the requirements are still in effect, because they have conflated CMMC with the requirements that CMMC verifies, when you say CMMC is suspended, they think the requirements are suspended, and it’s all your fault for not helping to clarify what’s going on. So after this leadership team is long gone, people will be writing checks to the government for their noncompliance, and it’ll be your guys’s fault.
Anyways, side note: I can’t wait till the DOJ gets more serious about export control. I think that will be a fascinating day, when we start seeing a bunch of, I don’t know if it’s False Claims Act, probably False Claims Act cases around export control specifically.
Daniel: Now, that does carry a criminal weight, so I don’t know if it’s just going to be a civil investigation, or I don’t know.
Jacob: Yeah, but I can’t wait for that day. The worst thing possible for defense contractors would be if the Department of State and the Department of Justice get really excited about AI capabilities, because the only thing keeping people from getting snapped off by government fines is capacity. And if they got really excited about AI tools allowing them to prosecute these cases faster, good luck.
Bonus points as a way around the suspension
Jacob: Okay. If you guys missed the podcast on Thursday, you should check it out. It’s very interesting, because during the suspension the DoD has said, “Hey, all you thousands of companies that have done the correct thing and gone to get your Level 2, those efforts weren’t in vain,” and they’re going to somehow help those companies or make it worth their while, or something. They didn’t say what that was.
Well, if you remember a couple of years ago, the Army MAPS contracting vehicle, a $50 billion contracting vehicle, had an approach to CMMC self-assessments and CMMC certifications that is not in rulemaking. It’s not part of the CMMC program. It was done of their own volition, and they basically said, “We’re going to award increasing amounts of bonus points for people who have Level 1 or Level 2,” or whatever. That was their solution to the problem. You don’t have to have the certification in order to win work under this vehicle, but if you do have the certification, you’re going to get a bunch of extra bonus points, and you’re much more likely to win the work.
Apparently that came out, and then after that was done, the suspension happened. So a few contractors protested this idea, saying, “Well, you can’t require it at all because it was suspended.” And GAO said that is irrelevant. They were like, the suspension doesn’t matter. They can require it if they want to. It’s just bonus points, right? We talked to Eric Crusius, great contracts attorney, and got his insight into how they approach it. You can find the link to the protest in the show notes for that podcast on Thursday. It’s a very interesting read.
But is this a method? Does this open the door to give the DoD an out to not be the bad guy? Because they can just say, “Well, it’s not required for contract award, but you will get more bonus points if you have the certification versus if you have a Level 2 self-assessment.”
Daniel: That would effectively require it without requiring it, because everybody’s going to scramble for any edge they can get to win the work, but technically the DoD isn’t requiring it. So they’re not actually the bad guy, because you don’t have to have it in order to win the work. I think that would be a very convenient way for them to do it. You don’t do this thing where you have to have it after the contract is awarded, because what if you can’t get it, and then they have to go back, and then it’s just a bunch of bureaucracy to deal with.
Jacob: It’s a big risk. If you have it as a condition of contract award, the department is just too afraid to draw the line in the sand, because it’s politically unpopular, and so they just don’t want to do it. They just aren’t going to commit. So this gives them an in-between option here. I think it would be very convenient, and GAO says it’s fine, so they should just do that. I really wish they would, at least in this interim era that we’re living in, this suspension, this pause era. Give people that did the right thing some kind of edge, even for a few months. It’s been 103 days. That’s the only new policy question raised during this review and RFI period: what are you going to do to help the people who got the cert? Every other question that was raised during the RFI has already been addressed many, many times before. CUI marking and scoping and small business and different requirements, this and that, over and over and over again.
Will the RFI produce new answers?
Jacob: We’ve already covered all that stuff. So unless the DoD is going to do a 180 on policies they’ve had in place for ten or fifteen years, you’re probably not going to get a new answer. We used to talk about this while we were waiting for the CMMC rules to happen, because people said, “Well, we had this 2020 rule. They’re going through rulemaking. What if they come out with something crazy?” And the question we would always pose to people is: this is what they’ve said on this position or this issue for many years now. You can look in the rulemaking and see their rationale. This is what the public comments brought up to them. What do you think their answer is going to be? Do you think their answer is going to be different, or do you think it’s going to be the same? If you think it’s going to be the same, you already have your answer for what will be in the rule. If you think it will be different, why do you think that? What evidence do you have? How different will it be? You’re taking massive risk, because if you’re wrong, you’re just doing it off of speculation.
This is the same situation, right? They asked all these questions probably because the people in charge don’t know what they’re talking about, and so they got a bunch of responses. Do you think they’re going to change their policy? Maybe they talk a big game, but I don’t see them changing their policy based off of what’s already happened, which is why we’re now over a hundred days later and we don’t have any real answers. I mean, who knows? Maybe they come out and go, “We’re just going to throw ten or fifteen years of policy consistency in the trash.” But I don’t think that’s what they’re going to do.
Daniel: Hey, I’ve got a Polymarket for post-midterms on the announcement, so I’m doing pretty good right now, quite honestly. (laughter)
Chat questions: Brilliant at the Basics and assessment criteria
Jacob: Let’s see. Gil said, “Is chat working?” Chat seems to be working for us. Everybody else said that chat’s working okay, so maybe drop out and then come back in. All righty.
Nicholas: With the pause, have C3PAO assessments changed in how they’re actually being done? Are assessors unofficially asking about any “Brilliant at the Basics”-type things?
Jacob: Not in our experience.
Daniel: None. Yeah, that was brought up at CS5, like, “Hey, this might be something to look into or familiarize yourself with if you don’t know.” But it has no bearing on the CMMC CAP, the CMMC assessment process, or anything like that. We were talking to people, because it’s basically, collectively, twenty good ideas of cyber hygiene and cyber practice and resilience, but there’s no assessment criteria, right? So even if they wanted an assessment included in that, there really isn’t anything there, and there’s no contractual obligation to do them either. So yeah, that was talked about at CS5 with a bunch of assessors, but there has been no action in relation to building it in, and they actually can’t build it into their assessment scope.
Jacob: So yeah, they’re not requirements, so there’s nothing to assess. And even if you were to label them as requirements, it would be the same problem that you have with the NIST Cybersecurity Framework. Take any of the Brilliant at the Basics and ask yourself, “What questions would I ask to know if this thing is implemented?” There are no verification criteria for those things. So there is no way for you to have a meaningful assessment approach to it, because this isn’t SOC. This isn’t ISO. This isn’t “you hire an auditor and they just ask whatever questions are in their head about how to prove whether something is in there.” NIST has standardized the set of questions that you ask in order to know if something is implemented. That’s also very helpful, because a lot of those things could be highly automated to make assessments faster and cheaper.
But when we see this push from the DoD and from FedRAMP and from other people toward what I call NIST CSF-ish verification criteria, you’re going in the wrong direction. You’re blending the details together and saying, “Are you resilient?” And you’re like, “I don’t know. How do I know if somebody’s resilient or not? What’s a measure of resilience?” The questions that I would ask would be different from the questions that you would ask. You have to have a standardized set of verification procedures, but a lot of people don’t have formal engineering backgrounds, and so verification and testing is a foreign concept to them. If you were to go into any real, legitimate engineering firm and say, “We’re just going to vibe-verify whether or not this thing is going to work,” you’d get laughed out of the room. That’s not how it works. You have to have verification criteria for engineering requirements. It’s bad enough that 171 cuts all the meat off the bone when it comes to these detailed requirements, but as of right now, we have not heard of anybody asking about Brilliant at the Basics-type things. I don’t know, chat, what do you think? Let us know.
Daniel: It’s Cybersecurity Awareness Month. The DoD CIO LinkedIn page is talking about Brilliant at the Basics on their posts. They have yet to post about the details of the actual Brilliant at the Basics document. (laughter) So we’re running out of time before the end of the month. There are twenty things on there. I haven’t heard them talk about any of that stuff in detail. So I don’t know, maybe they’ve got something big planned for the 31st or something.
What’s the plan after 103 days?
Jacob: It would allow them to provide incentive without acknowledging how unprepared the majority of the DIB is. Yeah, that’s really the rub, right? When you create a program to make people prove that they have implemented the requirements, because you know that they have not implemented the requirements pursuant to their contracts, and then you suspend that program because no one has implemented the requirements, you know, what’s the plan, right?
Daniel: What’s the plan? That’s what the program was created for. There are levers and buttons to pull within the management and oversight of the program, to have pressure release valves to allow people more time. But instead, we just suspended the mechanism that was designed to prevent people from getting the data without being able to prove that they can protect it. So what’s the plan going to be, 103 days in? We don’t have a plan. We’ll see what the plan is, but that really kind of gets to the heart of it. And that’s probably what the armed services committees are going to ask whenever they come back into session: what’s the plan?
I brought this up in a conversation at CS5. I wish the DoD had made the rollout of DFARS 252.204-7019 and the submission of your SPRS entry, you know, in 2020, “This was your score in 2021,” and just ratcheted it up to the launch, if you will, of CMMC. Because then at that point they would have been like, “We gave you a runway to the runway of the phased rollout of CMMC,” and instead of people feeling, quote-unquote, whiplashed, right? If you’re looking at it from an optics perspective, that would have helped a lot. Now, we all know on this call that 7012 has been around for a decade at this point, and people should have been doing this a lot longer. The DoD basically would have been doing it out of the kindness of their heart, which they didn’t really have to do. But I really wish there had been that one element. If I were to go back and tell the DoD, “Hey, do this one thing in preparation for CMMC,” I wish they would have done something like that. But hey, here we are now.
Jacob: There you go.
Jim: The bonus points don’t help those of us who have been third-party assessed and work only on purchase orders from primes or subprimes. We don’t bid directly on contracts.
Jacob: So I think this is saying you get bonus points for work underneath these vehicles if you have the cert. That’s true. Which sort of, we’ve talked about this issue a lot. All of these policy issues are the DoD talking to primes, but the majority, three-quarters, of the DIB are not primes. They are subcontractors. And so a lot of the issues that people have are raised by companies that the DoD isn’t interacting with directly. They’re working with them via proxy. That gap is a big problem, because the people in the middle, between the subs and the DoD, are the ones that have privity of contract with the subcontractors.
We saw for years, before CMMC was ever a final rule, some primes had minimum thresholds that your score needed to be before they would give you the work. Totally of their own volition. I was going back and forth with some people in the subreddit the other day, because they were like, “Some of the primes are requiring a Level 2, but DoD says that they can’t.” And they’re like, “No, no, no. The DoD says the contract workforce can’t require a Level 2 in DFARS 7021.” The DoD has no control over what the primes require or not.
That’s also one of the big questions that I have moving forward. There was no such thing as a November deadline for everyone to achieve Level 2 certification in DoD policy, ever. That was when Level 2 certification requirements would begin showing up in new solicitations. So the primes turned around and said, “All of you need to go have your Level 2 by November or we’re not working with you anymore.” The primes turned it into a deadline for everybody, and everybody freaked out. So if the DoD comes out and says we’re extending the deadline to November of 2027, the primes are going to turn around and say, “Get it by November of 2027 or else.” So if you have no ability to control that behavior from the primes around converting start times into deadlines, what’s the plan? What’s the plan?
Daniel: I don’t know. But apparently that’s what they’re working very hard to figure out while we all wait. If the phase rollout reboots, we’re just stuck in that same loop, right?
Jacob: I’ve said this before, but the story of DoD cybersecurity requirements and contracts, and CMMC, really is the story of endless concessions to industry to give more and more time to get ready. You have to go back and watch the history of CMMC video. The original DFARS 7012 came out in 2013. It required controls from 800-53, a larger set of requirements than what ended up being required in 2016 through 800-171. So you had a smaller set of requirements three years after DFARS 7012 was originally created. When DFARS 7012 came out in 2016 with that new set of requirements, everybody said, “We need more time to implement.” And the DoD was like, “Well, that doesn’t make a lot of sense, because this is actually less stuff than you’ve had for the last three years.” But then they gave everybody fourteen extra months to implement it. And that was December of 2017. And then you just go from there. It’s extension after extension, over and over and over again. It’s crazy.
Smith: We have an RFP from the Army Corps that requires a Level 2 self-assessment, with a couple of pages of CUI for critical infrastructure. If we redact or remove those pages, can we distribute to Level 1 subs for a bit?
Jacob: That’s a great question. If you remove the parts of the information that are CUI and there’s no CUI involved, then it’s FCI. Everything downstream from there would be FCI. Now, I think the problem that you’re going to have is you’re asking your government customer, and your government customer doesn’t know. This is where the language in DFARS 7012 paragraph (m) kicks in, where it says the contractor determines whether the data retains its identity as covered defense information. So if you have determined that it is no longer covered defense information, then that clause no longer applies and you don’t have to flow it down. You are making that decision, so the liability is on you, but you also have the ability to make that decision. So yep, if there’s no CUI, there’s no CUI. If you ask your government customer, you’re probably not going to get the answer that you want, because they don’t know.
I see this happen all the time between program managers at major prime contractors talking to their government customer. They say, “Is this information CUI?” And what the government customer says is, “All the data we sent to you is CUI,” which is true. Almost all these tech data packages that they send over, on space systems, satellite systems, weapon systems, air systems, blah blah blah, almost all that data is going to be CUI. So the government tells RTX or Northrop, “Everything I just gave you is CUI.” But Northrop is asking, “Is what I give to my sub CUI?” The government doesn’t know. It’s up to Northrop to make that decision. And what do they say? They take the risk-averse decision. They say, “We assume it’s all CUI. Go do these requirements.”
Daniel: Which I’ve seen people come back with the stance of telling the DoD, or whoever it might be, “Hey, these two or three pages are marked CUI. I’m not flowing down those pages to noncompliant subcontractors, but I am flowing down the rest of it as FCI.” And what I’ve seen is, if you come back with that warm blanket of “I’ve acknowledged CUI, I’m handling it correctly, the rest is FCI,” then you’re just leaning into a confirmation rather than a question. If you can word your emails into “Yeah, thumbs up, that’s great,” versus “Is this CUI?”, it’s going to go a lot better, because the minute you ask, to Jacob’s point, they say, “Oh, we only marked two pages. We meant to mark all twenty pages or all fifty pages.” And now you’re like, “Well, mother trucker, what am I supposed to do with this now?” So you do what the lawyers do: don’t ever ask a question that you don’t already know the answer to. We’ve said it a million times, the murkiness and gray area in government policy gives you the flexibility to really help yourself and your subcontractors out quite a bit.
Jacob: Every time you ask the government to clarify, you typically get more constrained rather than less constrained.
Brad: A cyber framework assessment that would be relevant to assurance requires an assessor with the ability to put the assessed organization’s business operations and practices in context relative to the CSF requirements.
Jacob: Sounds great. In my opinion, it’s a terrible idea. There are no standardized verification criteria for that framework. Therefore, it is a terrible idea to throw that out over the wall into the ecosystem and say, “You’re going to get assessed against a thing that we can’t tell you how you would know if you’ve done it or not.” It’s not a good idea.
The CSF is a great construct for understanding the 800-53 catalog, because the 800-53 catalog is written to be a dictionary. It is written to be a toolbox. It is written to be an encyclopedia. It is not written to be a front end. That makes it kind of difficult to work with, because it isn’t written in a way that has nice colors and easy-to-understand labels. It’s a dictionary. So you need a phrase book in order to understand the dictionary while you’re trying to travel around. You know what I’m saying? So the CSF is a front end to the 800-53 back end. The mistake is that people go, “Well, this summary thing that describes seventeen different 800-53 controls sounds nice, so just assess against this thing that sounds nice.” When you do that, you are committing the same flaw that NIST has committed, that NARA has committed, that the DoD has committed. You are assuming that the people who are reading what you are reading understand what you understand about technology and security and networking and all this stuff. They don’t. And so in order to help them, you have to put more detail into the requirements, not less.
Besides the fact that what you’re talking about here, Brad, with an organization’s business operations and practices, is wildly outside of the scope of what’s going on with the current requirements. CMMC 1.0 already tried this. They already tried to go into how you run your business and maturity and all those things, and they got smacked for it. So careful with the scope creep, careful with the allure of the easy-to-use CSF front end. Very valuable. Should be integrated more. Should be built out more. Totally agree. Not a good place to actually hold people’s feet to the fire on “are you doing this or not,” because there just isn’t enough verification within how those categories are written in the CSF, in my opinion.
Brad (follow-up): An assessor solely capable in technical assessment of control implementation would struggle with the CSF, which is why the 800-171 assessment objectives mapped to the CSF could be a solution.
Jacob: NIST has already mapped 800-171 to the CSF. 800-171 Rev 2: go to the page on NIST. They have a crosswalk to the CSF. They’ve already done it. If it helps people understand 800-171 better, wonderful. But verifying against the NIST CSF is not a good idea, in my opinion, because it’s more nebulous. In the same way, the Brilliant at the Basics document has the same problem. If you look in detail at what they’ve written in Brilliant at the Basics, at a high level it all sounds great. You’re like, “Yeah, sure.” But then when you read the explanations, there are compound requirements. There are completely different things included in the description of the original basic practice that they’re describing. It’s a nightmare for trying to verify what’s going on, because, like Jason and I were joking on the podcast, what does “your estate” mean in terms of your scope? I’ve never heard of this before in my life. What does that mean? There’s no definition. Might mean something different to you versus me.
Clarity is kindness, everybody, and that means details. That means details. We, as a community, as an industry, as a species, have got to get away from this idea that blurring things together makes it easier to understand for people who don’t know what you’re talking about to begin with. It makes it easier for you. It makes it easier for us. It doesn’t make it easier on the contractors. The paradox is, of course, there’s going to be more stuff to read and understand, but that’s a different story.
MITRE ATT&CK, CISA advisories, and the case for more detail
Daniel: You gave a talk at CS5 around the MITRE framework mapped to 171, and you ended up in your talk talking about 171 mappings. MITRE even has an explorer where they map these attack vectors to 800-53 requirements. Maybe CMMC just goes and points to 53 moving forward. Maybe 7012 gets updated to go to 53 instead of 171. Wouldn’t that be ironic?
Jacob: So, you know, Ron Ross said before he retired that 800-171 is moving back toward 800-53. If you listen to the many hours of conversations we had with Ron on the podcast, he acknowledged that taking the 800-53 requirements in the Moderate baseline and just absolutely shredding the details down to single-sentence descriptions of an outcome was a mistake, because people don’t know what you’re talking about. People don’t know what you’re talking about.
It’s like I talked about in that presentation. There was recently a CISA advisory that went out earlier this month, if everybody saw it, on Chinese APTs targeting critical infrastructure to steal sensitive data. Couldn’t have written it better myself. They map fifteen techniques from the MITRE ATT&CK framework that these Chinese advanced actors are using. How do they exploit? How do they get on? How do they stay on? How do they look around? How do they expand their footprint? How do they get the data out? These are known techniques that the bad guys use. Not everything is some spooky, alien-level zero-day for which there is no detection, no mitigation, nothing that you could do, and you’re completely helpless. A lot of what they do is perfectly normal stuff that is easily mitigated. People have this obsession with zero-days and advanced actors, like everything they do is a zero-day. Some subtlety, I beg of you, please, some subtlety for a moment, everyone.
The argument that you are too small to be a target is not a good one, because no one is too small to be a target if you’ve got the data that they care about. What is true is that advanced actors are probably not willing to burn multimillion-dollar zero-day super weapons on inconsequential targets. If you look far enough back on my LinkedIn, you’ll see my work history and where I used to work when I was active duty. I know a lot about how nation states operate on these networks, and I know a lot about how you can get the clearance to run an operation, allegedly, on networks. And you know what’s really hard to do? Run-of-the-mill targets don’t get approved for the super-crazy, alien-level, spooky capabilities. So if you are doing pretty normal things, it can stop even a nation-state in its tracks pretty well.
Anyways, those techniques mapped to forty-three controls from 800-53. A significant number of them are present in 800-171 and in 800-172 to some degree, but there is much, much more that has been cut out of those requirements that’s not in there. And there are many more things that are not in the Moderate baseline that could be used to amplify and enhance those requirements. So when you hear people like the DoD say, “We’re going to review the requirements,” you’re like, “So they’re going to go up, right?” In what possible way could you cut out less?
We won’t get into the details of the talk here, but a good example: go look at the 800-53 control SI-4, information system monitoring. It is the number one most commonly mapped control to these MITRE ATT&CK techniques used by real-world bad guys, because even when you can’t stop what they’re doing, for various reasons, you can almost always detect it, and you can almost always detect it after they’ve done it, and you get an alert or a bulletin from CISA or the FBI or whoever. You can’t detect it if you don’t have information system monitoring turned on. Now, when you look at the 800-53 control for SI-4, it has multiple parts to it, A through G. Only A is present in 800-171. Only the first part.
So if we were going to do a risk-based take on what the requirements should be, the number one thing that MITRE says will help against even the most advanced actors shouldn’t have any parts removed from it. That’s ridiculous. But if you go parts A through G, get ready for those assessment objective numbers to go through the freaking roof, because that’s just the way it’s going to work. Anyways, I won’t rehash the entire talk here. We’ll probably do a video on it in the future. But everybody says, even the DoD, everybody says we need to do a risk-based approach to the requirements. In what universe would a risk-based approach, a sober look at these current requirements, give you a baseline that is smaller than the one that you’ve had for the last ten years? That’s not what that means. (laughter)
Daniel: The only answer is the DIB just wouldn’t get the data. If they want to go that far, it’s going to come with more requirements. There’s just no way around it, really.
Portion marking and FAR CUI
Jim: What happened to government portion marking?
Jacob: I think experiences differ based on your customer, based on the planetary alignments, based on the weather, about whether you see marked data or not, the methods that they’re using to mark. Some people see portion markings, some people don’t. The governance over the CUI program within the department is garbage and terrible and inconsistent. And unfortunately, because of mismanagement by the Under Secretary for I&S and the people in charge of the DoD CUI program, we’re over here reviewing the CMMC program, which is completely different. It doesn’t even belong to that under secretary, and they have no authority or ability to influence the governance of the CUI program. I don’t know. What have you seen lately, Daniel, in terms of people with their marking? Because we run everybody through a discovery that rips through their network to try to find this information.
Daniel: I have seen very little of it, because they’re taking on too much risk by putting little paragraph statements or portion markings on pages. So they will just label the whole thing as controlled and classified. They’ll put a cover sheet on it and say, “The contents of this is controlled and classified information.” No portion marking. And we just see it all the time. Everyone’s trying to cover themselves in any fashion possible. And so it sucks, because when you get this down to the primes or the subs, something that could have not had to be marked CUI is now CUI, and now that’s limiting to the subs being able to access that information, because maybe they’re not compliant. Maybe they can’t self-attest to an 88 minimum requirement, right? Maybe they can’t do that kind of stuff. And so it ends up causing this immense amount of slowdown that brings way more organizations in scope of CMMC Level 2 requirements than it even should. That’s the number one problem, I think, that NDIA brought up in a lot of the responses to the CMMC pause RFI: “Hey, mark things appropriately, and I don’t think we’re going to have this problem as much.” That was kind of the general rhetoric presented in those responses, and they’re not wrong.
Jacob: I have a sinking feeling that the FAR CUI rule is going to be feast or famine, because that standardized form that they use to indicate whether there is CUI involved or not, if used appropriately, will be the single greatest development in this space in the last fifteen years, because it will eliminate the question of whether you have CUI or not. However, the behavior of the system is lazy and risk averse, and so it’s much easier for them to literally now check the box and say “this involves CUI,” and then you’ve got it in black and white that you have CUI. So it could be amazing. It could make things permanently worse. I don’t know. What do you guys think, chat, on this Friday? Do you think it’s going to be the thing that saves everybody, or do you think it’s going to entrench in the system that everybody has CUI?
Daniel: It’s just going to be a catchall. That’s my prediction. I hope not, but I can’t imagine changing that much human behavior all at one time.
Miss W: Besides FIPS not being properly implemented and lacking historical artifacts, which controls should I be wary of, and how do you mentally prepare for an audit like CMMC?
Jacob: My first recommendation for the list of things to watch out for: if you go to the DIBCAC website, they have a PowerPoint on there of the top ten most common “other than satisfied” requirements. They haven’t updated it in a while, which makes me think that they haven’t changed. That would give you a pretty good list of the things that DIBCAC commonly finds that people aren’t implementing well, because those are the things that are often marked as other than satisfied. So that’s a good place to start, just as a general rule of thumb. What do you think, Daniel?
Daniel: The things that we see over and over and over again, and I’m actually pulling up, this was an old blog by Amira that was posted of what those top ten were. FIPS, multifactor, identify, report, and correct system flaws, risk assessments, scan for vulnerabilities, review and update logged events, audit logging process failure alerts, review of audit logs, incident response capability, and then maintaining a baseline configuration. I have a feeling those have not drifted much.
What I will say that we see all of the time, and again, I go back to these not to sidestep the question, but because it’s right in front of our face: FedRAMP requirements of cloud providers, documentation of CUI asset types or general CMMC asset types in general. So if you have a CRM, make sure you can defend it, right? That’s kind of the wild card of the bunch. Security protection assets, all of that. Having a data flow, controlling the flow of CUI in your organization. What does that look like? When you look at the technical controls, especially if you’re doing an enclave, it’s a relatively easy deployment or something relatively easy to fix. A lot of what people see on this screen, they might pay an MSP or MSSP to take care of for them, right? I look at maintaining baseline configuration, audit review, scanning for vulnerabilities, multifactor enforcement, FIPS-validated cryptography. If you’re a small or medium business, you’re probably outsourcing some of these requirements.
And the question always comes back to: do you have a customer responsibility matrix mapped to the assessment objective level? And can your provider provide evidence that they’re actually doing these things? The number of conversations I have, and they have not stopped during the pause, of an MSP or an MSSP telling them they were good, and they spent thirty minutes popping the hood and realized, “Oh my gosh, they haven’t done any of this.” Or the one CMMC guy they had on their staff left six months ago, and nobody told the client, and nobody picked up the…
Jacob: Yeah, we’ve had that call before.
Daniel: Yeah. And it’s one of those things where they just don’t tell you unless you are checking their math. And listen, I know a ton of great MSPs. A lot of them were at CS5. Check out the MSPCollective.org website. That’ll list certified MSPs that are a really good starting point to talk to, but do some due diligence. You have to actually pop the hood and verify for yourself well before going into an assessment. And if you are an assessor, verify that if they have an MSP, they have an SRM, because you’re almost guaranteed to fail if you have an MSP and cannot produce that document. I don’t see how you could win, how you could pass.
Brad: I always thought an unending stream of 800-171 control implementation assessments was a good path toward burning out assessors.
Jacob: Yeah. Automate it. Automate them. You don’t need as many assessors. More assessors work more efficiently. They can run assessments faster. Automate them. It drives me insane, especially, not saying Brad’s saying this, but especially around the time of the suspension, all of these CMMC haters just came out of the woodwork and went, “Oh my God, an assessment every three years is so archaic. Human-run assessments is so stone age. Get out of the Jurassic and get with the times and automate it.” And I’m like, then automate it. No one’s stopping you. Just go automate it.
Daniel: The DoD has had STIGs for 800-53 control assessments for twenty years. Be my guest, please. It would be wonderful. You know the number of people we could take off of our compliance engineering team who have to handhold people through assessments? They could get back to the work of operating, maintaining, migrating, doing all the other stuff instead of having to babysit people through assessments. Sounds great to me. I’m sure they would love it.
Jacob: I’m sure they would love it. All these people who are like, “The clankers are going to do this, and oh, the AI is going to automate everything.” Then do it. No one’s stopping you. Go do it. What’s the problem? Go vibe-code up a bunch of agents that can run the assessments. DoD, why don’t you guys go do it? It’s been a hundred days. Spin up a hundred thousand agents. Let them loose. Let them swarm out there through the DIB.
Daniel: We’re already swarming through PHI and bank accounts and everything else these days. (laughter)
Jacob: Yeah, let them swarm through everybody’s CUI enclaves on a continuous 24/7 basis. You don’t have to pay them anything, because they’re a bunch of agents, right? And then see what happens whenever you have an assessment every three minutes instead of every three years, because nobody’s going to be able to pass. (laughter)
Daniel: But yeah, sure, it’s boring to run through the same stuff over and over again, so automate it. Before AI was ever a thing, the classic golden rule was if you have to do something more than once, just script it. So just automate it, please. I know none of the haters watch this stream. Be my freaking guest. If you guys want to go out there and automate the assessments, go do it, because then the assessment capacity argument is even easier, because all of the assessors that you have now are able to run and manage and oversee more assessments. Perfect. Everybody wins.
Jacob: Anyway, I was just about to say they’re going to have to come up with something, because we know a ton of C3PAOs that have had to let people go recently, in the past few weeks. So there’s going to have to be some kind of pickup if certifications are going to be requirements moving forward. So maybe agents are the way. Who knows?
Daniel: It’s still early data. We only have two real data points since the suspension, but it ain’t looking great. You’ll have to stick around, like and subscribe, follow us at summit7.us for the corresponding blog. But the initial set of data for increased assessment capacity in the ecosystem has slowed down for the first time in Phase 1, only since the suspension. If you look at the cone of possible growth, compounding being the most optimized, idealistic version, a linear growth rate being the most conservative version, you have this cone of probability of where you would end up over time. Since the suspension, the growth line is below the linear projection for the first time since Phase 1. So you didn’t have an assessor capacity issue before the suspension, but you do now. So I hope you guys got a good idea, because it’s all your fault. (laughter)
Sean: If DoD wants to place emphasis on resiliency or availability, how will they use 171? Does it only cover confidentiality?
Jacob: They would have to bolt more stuff on top of 800-171. They would have to take the 800-171 baseline, take all of the stuff that was tailored out of that baseline, and put it back in. If you remember far enough back to CMMC 1.0, that is exactly what CMMC 1.0 attempted to do. You can even find the Carnegie Mellon blog from their Software Engineering Institute where they talked about opening the aperture of 800-171 beyond just data confidentiality.
There are holes in the baseline on purpose, because NIST said, “We’re only going to take things from the Moderate baseline, which at the time was about 260 800-53 controls. We’re only going to include those controls that are relevant to the confidentiality of CUI, in their opinion.” And then they were only going to take the parts of those controls that were relevant to protecting the confidentiality of CUI, in their opinion. So you are many, many, many steps away from the full CIA triad version of the 800-53 baseline. So if they want to increase resiliency by focusing on availability and integrity instead of just confidentiality, those requirement numbers are going way up. Those assessment objectives are going through the freaking roof, which, you know, sounds great from a security nerd perspective. Jack me up. Just let her rip. Throw them all in there. Let’s go back to the 800-53 Moderate baseline. Do all 260 of them. I think that sounds great. What we all know is that’s not going to sound great to anybody else.
Daniel: No one will like that. Or their pocketbook, quite honestly. You start looking at that and you’re like, if we’re trying to eliminate the cost burden for small businesses, if you ratchet up those requirements and don’t provide funding, where’s the money, or solutions hosted by the DoD or another federal agency? What are you going to do?
Jacob: Yeah. So are you just going to increase taxes and not close any of the loopholes? Be like, “Sure, in that case, just tell everybody that they have to have a billion-dollar security budget every year, and you’re never going to ask them to prove it. So it doesn’t really matter what you ask them to do.”
Daniel: Justifying the suspension and a review of requirements based off cost, burden, and impact, to then turn around and say that we need to do a risk-based look to expand the requirements, is a contradiction. And I don’t know how they’re going to make it square up, which is why we’re all waiting with excitement and bated breath to see what their answer is.
Jacob: Here’s the thing. If old-school CMMC 1.0 Level 3 came back and backups were back in scope, I don’t think anybody would have a problem with that.
Daniel: I don’t think so. I think there are some low-hanging fruit that they could potentially incorporate back without significant cost, but it would be a cost.
Jacob: Yeah. I think there’s stuff that they could tailor out of the baseline. I’m not out here saying that the baseline is perfect. I will grant to everybody that the version of the 800-171 baseline as it exists probably isn’t the ideal version of what it should be. But if everybody says it needs to change, it’s on you to say what needs to change. What are you taking out? What are you putting back in? That tailoring process is not a trivial thing to do. And what are we going for here? Are we going for resiliency and security? Are we going for cost? Are we going for risk? You might get different answers based on which criteria you use to do your tailoring. I just don’t see a way that you end up with a lesser, smaller, cheaper baseline in almost any of those situations.
Hotline queue questions
Jacob: All righty. A couple of queued questions here. Chat’s been on fire today. We’ve got ten minutes left before we get to the queue. This is awesome.
Would switching providers for GCC High licensing trigger a new audit?
Daniel: No, because you’re not switching from a not-met control to a met control. You’re just switching where you’re getting the licensing from. That is a non-triggering event, if you will. That’s just picking up a new provider.
Can we keep using unsupported Windows 10 computers on manufacturing equipment and still meet CMMC Level 2?
Daniel: You can. Between a combination of specialized assets, and there’s no requirement in Rev 2 to be on a modern or supported version of an operating system. Now, there are things like risk assessments you have to do. You might want to isolate that thing off the network, and so forth. But yes, you can still use Windows 10 computers until our good friend Rev 3 starts knocking on the door. And if they change, and this will be interesting, Jacob, do you think they will change scoping guidance at all?
Jacob: This is the crazy part. This is what we talked about in the episodes two weeks ago on Brilliant at the Basics. We talked about this specifically. There’s no scoping guidance for Brilliant at the Basics. They don’t seem like they’re big fans of how the CMMC program came together, and they label it all as red tape and bureaucracy. The problem is that the CMMC program is responsible for the most detailed scoping guidance that the DIB has ever gotten. And so if you shred the scoping guide for CMMC, you go back to the single-sentence scoping guide that NIST provides in 800-171, which is just everything. There are no specialized assets. There are no contractor risk-managed assets. There are no out-of-scope assets. There are no designations around those things outside of the CMMC scoping guide. So if you rip up the program, or if you rip up the scoping guide, it makes it worse, because you don’t have these outs to save yourself money.
Daniel: It’s kind of crazy, right? The FAR CUI rule doesn’t have a scoping guide. I want to say I even saw a public comment addressing that. I think maybe Allison posted it: “Hey, is this everything, or can we have some sort of scoping guidance?” And I’m assuming, just like they left assessments up to each agency, that they think every agency will basically build their own scoping guidance around Rev 3, which, by the way, I have a feeling none of them have been given a heads-up about coming down the pipe. So that’s going to be really interesting, quite honestly.
Jacob: So much for all that harmonization. Am I right?
Daniel: Yeah, man.
Jacob: Spicy.
How can a prime contractor verify a subcontractor’s SPRS score if primes can’t access those records directly?
Daniel: Welcome to a multimillion-dollar industry of what is supply chain risk software. You’ve probably heard of tools like Exostar before, or other tools like that. I think FutureFeed now has a module for something like this, where they’re basically having the subs submit their score and attest to it, sometimes having to provide evidence depending on where they sit. But yeah, primes do not have access to SPRS. In no world are they ever going to have access to SPRS. They’re going to have to collect them manually. There was conversation at some point about trying to build a centralized validation of CMMC certifications, which would have been really nice just to validate, “Hey, does contractor ABC really have this?” and then have the assessor basically validate that on some sort of public forum. But I haven’t heard any update on that in a very long time. So, long story short, primes don’t have access. They’ll have to trust the supplier or the sub until maybe there’s some kind of third-party independent verification, almost like a Credly when you get a certification or something like that.
Jacob: The Chinese MSS could do the funniest thing ever right now and just do a public-facing web tool where you could query defense contractors. They can tell you what the score is.
Daniel: Oh man, I would not run it past them.
Jacob: Oh man, they could have FedRAMP Moderate equivalent data centers in China. I’m not saying that your score is CUI, but just making a little bit of an ultimate April Fools’ prank if you’re out there listening. Anyhow, everybody.
Jim: What about the cost for a small business with 25 endpoints to purchase software seats, when many vendors have 50 to 150 minimum seat requirements plus an additional annual charge of $10,000 to $20,000 per year?
Daniel: Depending on the software stack, like through the AOSG program with Microsoft, we can sell single licenses. We don’t have to sell minimums of fifty or a hundred or five hundred in the Microsoft world of things. But there are some vendors where, to get their GovCloud version, you have to have minimum seat counts. They’re just not going to let you in the door without it.
Jacob: Yep. I remember when I was at the NIST Manufacturing Extension Partnership program, our job was to go around to small defense contractors in Southern California and be like, “Do you know what DFARS 7012 is? Have you ever looked into 171? You’re telling the government you’re doing this. We’re here to help explain it at no cost to you.” It’s what I did for a couple of years, and it was very interesting because you talk to different companies every week. I remember I called Darktrace, because Darktrace is super cool. And I was like, “We’ve got a real problem on our hands, guys. We have this massive critical infrastructure sector that needs a lot of help, and you guys are super badass. So do you want to help us out?” And they were like, “How big are these companies?” I was like, “Sometimes it might be ten people.” They hung up on me. I tried again like six months later, and they hung up on me again. So, you know, I’m glad that you guys have your logo on the back of an F1 car, but that wasn’t necessary.
Daniel: They didn’t know. The SMB might not be their market.
Jacob: No, no. And they didn’t even bother telling me it wasn’t their market. They just were like, “Click. See you.” (laughter) So, there you go. All righty.
Does FedRAMP High automatically satisfy ITAR requirements, including restrictions on foreign person access? We’ve talked about FedRAMP Moderate a lot. What about FedRAMP High?
Daniel: It does not. There is no US person, US citizen data sovereignty situation in play with FedRAMP High. It’s a set of baselines and standards that they’re assessed against, not personnel and location. That’s a little problematic. I run into people all the time who are like, “Oh, I’m using a FedRAMP High service. Oh, I’m using a FedRAMP Moderate service. Therefore, I have US persons, or I can store export control.” A lot of times they’ll say they can store export control because they don’t really know what the ITAR requirements are. They just assume FedRAMP, federal equals federal, equals all federal data, is kind of the logic train followed there.
In reality, no. You typically have to get completely separate support contracts, and have specific data stores that you’ve established in their data centers where export control data will reside within the United States or a friendly country, with appropriate encryption in play and all of that good jazz. I wish there was some more clarity around that. I actually went to the FedRAMP FAQs a couple of weeks ago, and they don’t really talk about personnel or export or anything like that. It might not be a bad FAQ for them to add, just for some clarity’s sake.
Wrap-up
Jacob: Yeah, good point. All right, everybody, it’s an hour. It always flies by. This is the last weekend before Halloween, so make sure you’re out there decorating for the kids and keeping it alive. It’s the last Halloween on a Saturday for the next eleven years, so make it count. It’s also daylight saving time, so get the big candy bars.
Daniel: Yeah, ideal times. There won’t be a Halloween like this for another eleven years. So there you go.
Jacob: Okay. Make sure you like and subscribe. We are planning on a bunch of big things happening before the end of the year in the regulatory space: CIRCIA, FAR CUI, whatever the heck the DoD decides to do, all kinds of fun stuff. You can follow us on LinkedIn. You can follow Summit 7 on LinkedIn. Summit7.us. You go to cuihotline.org and fill out the form. You call the number and leave a voicemail. If we get to your question or comment that you leave after this stream is over, then we’ll add it to the queue like we always do. We’ve got quite a large queue at this point, so tons of stuff to burn through.
And there you go. Thanks, everybody, for stopping by. I think that was the longest time period that we’ve gone on the stream before getting to the queue, because of how active chat was. So props to you guys. Always fun hanging out with you every Friday, and we’ll see you next week. See y’all.
Contact
Speak With Our Team
Our team of compliance and cybersecurity experts are on standby and ready to help. We’ll walk you through what you need and what to expect.
