The 2018 cyber strategy moved federal supplier security into the contract file

The September 2018 National Cyber Strategy promised streamlined vendor exclusion, testing on contractor systems, consistent contract clauses and a government that met its own standards. Each was a procurement job with its own instrument, and each moved at its own speed.

Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.

Two-minute summaryThe dek and the first sentence of each section, quoted from the article

The September 2018 National Cyber Strategy promised streamlined vendor exclusion, testing on contractor systems, consistent contract clauses and a government that met its own standards. Each was a procurement job with its own instrument, and each moved at its own speed.

  1. The week On Thursday, September 20, 2018, the White House released a National Cyber Strategy.
  2. Exclusion: who may sell to an agency The strategy’s supply chain paragraph made a specific complaint about the federal acquisition system.
  3. Access: what the government may do on a supplier’s systems The contractor paragraph went further than any other part of the federal section.
  4. Uniformity: one clause instead of many The strategy said the administration would support “adoption of consolidated acquisition strategies to improve cybersecurity and reduce overhead costs associated with using inconsistent contract provisions across the Federal Government.”
  5. Example: the government’s own systems and the projects it funds The fifth priority action turned the demand around.
  6. How the four jobs reach banks, health systems and state agencies The federal section of the strategy addressed federal agencies and their contractors.
  7. Objections: a strategy is not a rule The strongest objection is that a strategy changes nothing by itself.
  8. What a program office or a supplier can do with this today The four jobs give a simple way to audit an agency’s supplier security work, or a supplier’s readiness for it.
  9. With hindsight The exclusion job got its statute that December.

The week

On Thursday, September 20, 2018, the White House released a National Cyber Strategy. A fact sheet issued the same day called it the first fully articulated National Cyber Strategy released in 15 years. The fact sheet said the strategy built on the President’s executive order “Strengthening the Cybersecurity of Federal Networks and Critical Infrastructure,” signed in May 2017, which the strategy identifies as Executive Order 13800. The strategy document itself, dated September 2018, said that the work performed and reports created in response to that order laid the groundwork for it.

The strategy was organized around four pillars, following the National Security Strategy. The first was to protect the American people, the homeland and the American way of life. The second was to promote American prosperity, the third to preserve peace through strength and the fourth to advance American influence. The document said the National Security Council staff would coordinate with departments, agencies and the Office of Management and Budget (OMB) on an appropriate resource plan to implement it.

The first pillar opened with securing federal networks and information, and listed five priority actions under that heading. The first was to further centralize management and oversight of federal civilian cybersecurity in the Department of Homeland Security (DHS), with the exception of national security systems and Defense Department and intelligence community systems. That included ensuring DHS had appropriate access to agency information systems for cybersecurity purposes and prioritizing the move of agencies to shared services under OMB oversight. The second was to align risk management and information technology activities: under Executive Order 13833, agency leaders would hold their chief information officers accountable for aligning cybersecurity risk decisions with IT budgeting and procurement decisions.

The third action was to improve federal supply chain risk management. The strategy said the administration would integrate supply chain risk management into agency procurement and risk management processes, improve information sharing about supply chain threats among agencies and create a supply chain risk assessment shared service. It would also address deficiencies in the federal acquisition system, “such as providing more streamlined authorities to exclude risky vendors, products, and services when justified.”

The fourth action was to strengthen federal contractor cybersecurity. The strategy said the federal government would be able to assess the security of its data by reviewing contractor risk management practices and by testing, hunting, sensoring and responding to incidents on contractor systems. Contracts with federal departments and agencies would be drafted to authorize those activities. It named contractors in the defense industrial base that research and develop key Defense Department systems as an acute concern. It said the administration would support consolidated acquisition strategies to reduce the overhead of inconsistent contract provisions across the government, and would work to ensure contractors received relevant threat and vulnerability information.

The fifth action was for the government to lead in best and innovative practices. Systems the government owns and operates would meet the standards and practices it recommends to industry, and projects that receive federal funding would have to meet them as well. The government would use its purchasing power to drive improvement across sectors. As an example, the strategy said the National Institute of Standards and Technology (NIST) would continue to standardize quantum-resistant public key algorithms, and that the government should adopt those standards quickly. Outside the federal section, the strategy listed seven key areas for reducing risk to critical infrastructure: national security, energy and power, banking and finance, health and safety, communications, information technology and transportation.

The strategy’s fact sheet summarized all four pillars on a single page. The strategy’s section on federal networks read at first like an internal security plan. Read closely, three of its five priority actions carried four procurement commitments. Each one promised to change who could sell to an agency, what an agency could do on a supplier’s systems, what agency contracts would say, or what conditions came attached to federal money. A security team could not deliver any of that alone. Each needed a legal authority, a contract clause or a funding condition, and each moved at its own speed.

The distinction matters for anyone who reads a strategy as a forecast of obligations. A security leader who sees “strengthen federal contractor cybersecurity” in a strategy will ask what controls are coming. A contracting officer will ask what clause, under what authority, starting when. A budget officer will ask who pays. The federal section answered the first question in general terms and left the other two to the agencies and to rulemaking. That gap between the security promise and the procurement instrument is where the work sat in 2018, and where it still sits for any agency that writes security requirements into what it buys.

Our argument is that the strategy’s supplier commitments are best read as four separate procurement jobs, each owned by a different part of an agency and each with a different instrument. We take them in turn:

  • Exclusion: the power to keep a risky vendor or product out of federal purchases, which needs a statute and a contracting process.
  • Access: the government’s ability to review, test and respond on a supplier’s own systems, which needs contract language the supplier signs.
  • Uniformity: one set of security clauses across agencies instead of many, which needs a governmentwide regulation.
  • Example: the government meeting its own standards and attaching them to the projects it funds, which needs budgets and grant conditions.

The four do not overlap. Exclusion decides who is in the supply chain. Access decides what the government can see once a supplier is in. Uniformity decides how the terms are written. Example decides whether the buyer holds itself to the same terms. An agency could make progress on one and none of the others. After the four parts, we weigh the main objections and set out what a program office or a supplier can do with this today.

Exclusion: who may sell to an agency

The strategy’s supply chain paragraph made a specific complaint about the federal acquisition system. It said agencies needed “more streamlined authorities to exclude risky vendors, products, and services when justified.” That sentence treated the existing authorities as a deficiency in the acquisition system. What it asked for was a way for an agency to say no to a supplier because of supply chain risk, and to make that decision hold across other agencies.

Exclusion is the bluntest procurement tool, and the hardest to use well. As we read federal acquisition, a contracting officer would struggle to reject an offer on a suspicion alone: awards rest on competition and on written evaluation criteria, and a disappointed offeror may challenge the decision. To exclude a supplier on security grounds, an agency needs an authority that says it may, a process that produces a record, and some way to share the decision so that the next agency does not buy the same product a week later. The strategy’s phrase “when justified” pointed at that record. An exclusion that cannot be explained will not survive a challenge.

The strategy paired exclusion with two other supply chain promises: better information sharing about supply chain threats among agencies, and a supply chain risk assessment shared service. Those belong in the same part, because they feed the exclusion decision. An agency cannot exclude what it does not know about, and a small agency cannot afford its own intelligence on every supplier. A shared assessment service would let many buyers rely on one analysis.

The strategy also said the effort would be synchronized with efforts to manage supply chain risk in the nation’s infrastructure. That clause linked the federal buyer’s problem to the private operators of critical infrastructure, who buy from the same suppliers. A product that a federal agency decides to exclude may also run in a bank’s data center or a hospital’s network. How, and whether, an exclusion decision would be shared with those operators was left open.

For an agency, this part lands in three offices. The acquisition office owns the process and the record. The chief information officer and the security team supply the technical judgment about what makes a product risky. General counsel decides whether the record is strong enough. In 2018 those three offices could each have pointed to the strategy and said the exclusion job belonged to someone else. The strategy did not say which office would own it inside an agency; it described a governmentwide gap.

State and local agencies face the same problem with fewer tools. Some buy through cooperative contracts or state schedules that another body negotiated. If a product on such a vehicle is later found risky, the agency that uses the vehicle may have no say in removing it. The federal strategy did not address state purchasing, but the logic carries: exclusion only works when it reaches every route by which a product can be bought.

Banks and health systems are not federal agencies, and the strategy did not bind them. Their own vendor management rules are a separate matter. The useful parallel is the record. A bank that drops a vendor on security grounds, or a health system that refuses a device, is making the same kind of decision. It holds up when the reasons are written down at the time, tied to a stated criterion and shared with the people who might buy the same thing elsewhere in the organization.

What a defensible exclusion record contains

The strategy did not say what an exclusion record should contain, and we do not claim it did. Working from the logic of the paragraph, we would expect a record that can survive review to answer five things. What is being excluded: a named supplier, a product line, a component or a service. Why: the specific risk, described well enough that someone outside the security team can follow it. On what evidence: the assessment, intelligence or test result relied on, and who produced it. What alternatives were considered: mitigation, a narrower restriction, a time limit. Who decided, and on what authority.

Each of those elements is cheap to write at the time and expensive to reconstruct later. An agency that decides informally to stop buying a product, without that record, may find that the next contracting officer buys it again because nothing in the file says not to. The shared assessment service the strategy proposed would have supplied the evidence element for agencies that could not produce it themselves. The other four elements still belong to the deciding agency.

Access: what the government may do on a supplier’s systems

The contractor paragraph went further than any other part of the federal section. It said the government would be able to assess the security of its data “by reviewing contractor risk management practices and adequately testing, hunting, sensoring, and responding to incidents on contractor systems.” It then said contracts with federal departments and agencies would be drafted to authorize such activities.

That is two different kinds of access in one sentence. Reviewing a contractor’s risk management practices is an assessment: the government looks at policies, evidence and perhaps an audit result. Testing, hunting, sensoring and responding on a contractor’s systems is operational: the government, or someone acting for it, runs tools on equipment the contractor owns, looks for intruders there and acts when it finds one. The first fits inside a normal contract. The second changes the relationship between buyer and supplier.

Operational access raises questions that an agency has to answer in the contract before it can rely on the access. The contract has to say which systems are in scope, whether only those that hold government data or everything connected to them. It has to say who runs the sensors and who sees what they collect, and what happens to the supplier’s other customers’ data, which may sit on the same systems. It has to say who pays for the work and for any outage it causes, and what happens when the hunt finds something the supplier did not report. The strategy did not answer those questions. It said the contracts would be drafted to authorize the activities, which places the answers in contract language.

The strategy named the defense industrial base as an acute concern, and singled out contractors responsible for researching and developing key systems fielded by the Defense Department. That made sense as a priority. It also showed where the access question would be hardest. A large defense contractor runs corporate networks that serve several programs and customers. A small subcontractor may run a few laptops and a cloud email account. The same clause would mean very different things to each.

Subcontractors complicate the access part further. The strategy spoke of contractor systems, and a prime contractor’s systems may be only part of the picture. Government data can pass to subcontractors and their suppliers. An access clause that reaches only the prime contractor leaves the rest of the chain outside the government’s view. Extending it down the chain means each prime has to pass the terms to its subcontractors and enforce them, which is a contracting burden of its own and one that falls hardest on the smallest firms.

For agency leaders, the access part belongs to the program office and the contracting officer as much as to the security team. The program office knows which supplier systems hold the agency’s data. The contracting officer writes the terms that let the agency see into them. The security team decides what it would do with the access. If any one of the three is missing, the clause either is not written or is written and never used.

For a supplier, including a health system or a university hospital that holds federal research awards, the access part is a negotiation. It is reasonable to ask what the government would run, on which systems, with what notice and under what confidentiality terms. It is also reasonable to expect that a supplier that refuses all access will find some contracts harder to win.

Five terms an access clause needs

An access clause that is meant to be used, as well as signed, has to settle at least five terms. These are our list, drawn from the questions above, not terms the strategy set out.

  • Scope: the systems covered, defined by the data they hold or the services they deliver, with a way to update the list as the supplier’s environment changes.
  • Method: what the government may do on those systems, from reviewing evidence to running scans to placing sensors, and who carries out each activity.
  • Notice and timing: how much warning the supplier gets for routine testing, and what changes during a suspected incident.
  • Data handling: what the government may collect, how it protects the supplier’s other customers’ data, and how long it keeps what it collects.
  • Cost and liability: who pays for the work, and who bears the cost of an outage or a false alarm caused by it.

Without the first term, the clause covers nothing in particular. Without the second, the supplier can argue that any specific activity was not agreed. Without the third and fourth, a supplier’s counsel will slow every request. Without the fifth, the program office will find that nobody budgeted for the access it negotiated.

Uniformity: one clause instead of many

The strategy said the administration would support “adoption of consolidated acquisition strategies to improve cybersecurity and reduce overhead costs associated with using inconsistent contract provisions across the Federal Government.” The phrase described a cost that suppliers to many agencies knew well. Each agency, and sometimes each program, wrote its own security clauses. A supplier selling the same service to five agencies could face five sets of terms, five reporting deadlines and five definitions of an incident.

Inconsistent clauses are an economic problem before they are a security problem. Every variation costs legal review, a compliance mapping and sometimes a separate operating process. That cost weighs more on a small supplier, which has fewer contracts to spread it across. If it decides not to bid, the agency has fewer offers, and the offers that remain price in the compliance cost.

Uniformity also has a security effect. When every contract describes incident reporting differently, a supplier facing a real incident has to check which clause applies to which data before it can report. That delay is the opposite of what the clauses were meant to achieve. A single clause, written once, lets the supplier build one process and test it.

The instrument for uniformity is the Federal Acquisition Regulation (FAR), which applies across executive agencies, and its agency supplements, such as the Defense Federal Acquisition Regulation Supplement (DFARS). A strategy cannot amend the FAR. It can only direct the people who do. So the uniformity promise depended on rulemaking, which runs on its own timetable with proposed rules, comment periods and final rules.

For a chief information officer, the practical point is that uniformity reduces the agency’s freedom to write its own terms. A CIO who wants a stricter reporting deadline for one high-risk system may find that a governmentwide clause sets the floor and that adding agency-specific terms brings back the cost the strategy wanted to remove. The trade-off is real and belongs in the acquisition strategy for each major system, which should record where the standard clause suffices and where a deviation is worth its price.

Banks and health systems face a version of the same trade-off with their own vendors. An institution with dozens of contract templates for technology suppliers carries the same overhead the strategy described. A standard security schedule, used across contracts and changed only by exception, is the private-sector equivalent of the consolidated acquisition strategy.

Where variation is worth keeping

Uniformity is not always the goal. Some variation exists for good reasons: a system that controls physical equipment needs different terms from an email service, and a contract for a classified program needs terms that an unclassified one does not. The strategy’s phrase was “inconsistent contract provisions,” not “all variation.” The distinction we would draw is between variation that tracks a real difference in risk and variation that exists because two offices wrote their clauses separately.

A simple test sorts the two. For each agency-specific clause, ask whether a supplier could meet the governmentwide clause and still leave the agency exposed to the risk the specific clause addresses. If yes, the variation is earning its cost. If no, it is the overhead the strategy described, and the agency pays for it in fewer bids and higher prices.

Example: the government’s own systems and the projects it funds

The fifth priority action turned the demand around. The strategy said the federal government would ensure that the systems it owns and operates meet the standards and cybersecurity best practices it recommends to industry. It added: “Projects that receive Federal funding must meet these standards as well.” It said the government would use its purchasing power to drive sector-wide improvement in products and services.

Those three sentences contain three separate commitments. The first is about the government’s own systems, and its instrument is the agency budget: meeting a standard costs money and staff time. The second is about grantees and other recipients of federal funds, and its instrument is the condition attached to the award. The third is about markets, and its instrument is buying volume: if the government only buys products with certain features, suppliers build those features, and other buyers benefit.

The second commitment reaches furthest beyond Washington. Federal funds flow to state and local agencies, universities, hospitals and research institutions. If the standards the government recommends to industry were attached to those funds, each recipient would need to show it met them. The strategy stated the principle but did not name the standards or the funding programs. That left the grant offices and program offices that write award terms to decide how far to go.

The strategy’s example of leadership was cryptography. It said NIST would continue to solicit, evaluate and standardize quantum-resistant public key algorithms, and that the United States should support “rapid adoption of these forthcoming NIST standards across government infrastructure.” That is a useful test case for the example part, because it shows the order of work. A standard has to exist before anyone can be held to it. Then the government has to adopt it in its own systems. Only then can it credibly ask suppliers and grantees to do the same.

For an agency, this part lands on the CIO and the budget office together. A commitment to meet the same standards the government recommends to industry is easy to write and expensive to keep. It means legacy systems that cannot meet the standard need a funded plan, an exception or a retirement date. For a grantee, the part lands on the research or program office that signs the award, which may not know that a security condition is in it.

The cryptography example also shows why the example job cannot wait for the other three. A supplier asked to adopt a new algorithm will reasonably ask whether the agency’s own systems can accept it. If the agency’s systems cannot, the supplier’s investment buys nothing until the agency catches up. The buyer’s readiness sets the pace for the whole chain, which is one more reason to start with the government’s own systems rather than end with them.

For banks and health systems, the example part has a direct analogue in third-party risk management. An institution that asks its suppliers for controls it does not apply to its own systems will find the request hard to defend in a negotiation and in front of an examiner. The government faced the same credibility problem, and the strategy acknowledged it.

Grant conditions as security controls

The sentence about federally funded projects deserves separate attention, because award conditions work differently from contract clauses. A contract buys a defined product or service; the government is the customer. A grant or cooperative agreement funds the recipient’s own program; the recipient runs it. A security condition in a contract describes how the supplier must protect the government’s data or systems. A security condition in an award describes how the recipient must run its own systems while spending federal money.

That difference changes who reads the condition. In a contract, the supplier’s contracts and security teams read the clause before signing. In an award, the condition may sit in general terms that a program or research office accepts without a security review. For a state agency receiving federal program funds, or a hospital with a federal research award, the practical step is to route each award’s terms through the security office before acceptance, and to keep a list of the security conditions in force across all awards.

How the four jobs reach banks, health systems and state agencies

The federal section of the strategy addressed federal agencies and their contractors. It did not regulate banks or hospitals. Its reach into those sectors came through three routes, and they map onto the four jobs differently.

The first route is direct: an institution that sells to or holds an award from a federal agency is on the supplier side of the jobs. A health system with federal research grants, a bank that provides services to an agency or a state agency that administers a federally funded program would face the access terms in its contracts and the security conditions in its awards. For those institutions, the exclusion job can also bite: a product they use to deliver a federal contract could become one they are no longer allowed to provide or use for that work.

The second route is by example. The strategy said the government would use its purchasing power to drive sector-wide improvement in products and services. If federal buyers insist on a feature, a supplier may build it into the product it sells to everyone. A bank or hospital buying the same product may then get a security feature it did not negotiate for. That benefit is real but uneven; it reaches only products that the government buys in volume.

The third route is by priority. Outside the federal section, the strategy listed seven key areas for reducing national risk, including banking and finance and health and safety. That listing did not create obligations. It signaled where federal engagement on critical infrastructure would be concentrated, which matters to institutions that work with their sector’s federal partners.

For a bank’s or health system’s third-party risk program, the four jobs offer a useful audit frame even without any federal connection. Exclusion asks whether the institution can stop buying a product across every business line once it decides to. Access asks whether its contracts let it see into a critical vendor’s environment when it matters. Uniformity asks how many different security schedules it signs and whether each variation earns its cost. Example asks whether it meets the controls it demands of vendors. A program that can answer those four questions with evidence is in a stronger position than one that can only show a vendor questionnaire.

For a state or local agency, the frame applies twice: once as a buyer from its own suppliers, and once as a recipient of federal funds that may carry conditions. The two roles can conflict. An agency may be asked to meet a federal standard in a funded program while its own contracts with the vendors that run that program say nothing about the standard. Matching the two is a contracting job, and it is easier before the next renewal than during an audit.

Objections: a strategy is not a rule

The strongest objection is that a strategy changes nothing by itself. It binds no contractor, amends no regulation and appropriates no money. That is true, and it is why we read the federal section as a list of jobs rather than a set of results. The value of the document, for an agency leader, was that it named the jobs and the order in which the administration said it would pursue them. The strategy itself said the National Security Council staff would coordinate a resource plan with OMB and agencies; it did not publish one.

A second objection is that supply chain and contractor security were already moving before September 2018, especially in defense contracting, and that the strategy described work already under way. The document did present itself as building on earlier work under Executive Order 13800. That does not reduce the usefulness of separating the four jobs. It means some of them started further along than others.

A third objection is that the four parts are really one: buy secure things from secure suppliers. We think the separation matters because the instruments differ. Exclusion needs a statute and a process. Access needs signed contract terms. Uniformity needs a governmentwide regulation. Example needs budgets and award conditions. An agency that tracks them as one item will report progress on whichever moves fastest and miss the others.

The separation also helps with accountability. When a supplier incident happens, the after-action question is which job failed. A risky product may have been bought because no exclusion process existed. The agency may have learned of the incident late because the contract gave it no access. The supplier may have misreported because three clauses conflicted. The agency may have demanded a control it did not meet itself, and lost the argument when the supplier pushed back. Each answer points to a different office and a different fix.

A fourth objection, from suppliers, is that more access and more conditions raise costs that the government ends up paying. That is also true. The strategy’s own uniformity promise was an attempt to limit those costs by removing the variation between agencies. The question for each contract is whether the security gained is worth the price, and that question belongs in the acquisition strategy, not in a security policy alone.

A fifth objection is that exclusion powers can be used for reasons other than security, such as trade policy or favoritism, and that a streamlined authority makes misuse easier. That risk is the reason the record matters. An exclusion that rests on a written, reviewable rationale is harder to misuse than one made informally, and the strategy’s own qualifier, “when justified,” points the same way.

What a program office or a supplier can do with this today

The four jobs give a simple way to audit an agency’s supplier security work, or a supplier’s readiness for it. For each, the question is who owns it, what instrument they use and what evidence shows it works.

  • Exclusion: name the office that would decide to stop buying a product for supply chain reasons, and test the process on one past case. Check whether the decision would reach every contract vehicle the agency uses, including cooperative and governmentwide contracts.
  • Access: list the supplier systems that hold agency data. For each, read what the contract lets the agency see, test or require during an incident. Where the answer is nothing, decide whether that is acceptable.
  • Uniformity: count the security clauses in use across the agency’s major contracts and the variations between them. Where an agency-specific clause exists, record why it was worth the cost.
  • Example: compare the standards the agency asks suppliers to meet with the ones its own systems meet. Where there is a gap, record the plan, the exception or the retirement date.

For suppliers, including health systems and banks with federal contracts or grants, the same four questions run in the other direction. A supplier should know whether a product it sells or uses has been excluded, what access it has agreed to give and whether it could deliver that access during an incident, which clauses it operates under and whether its processes match each one, and whether it meets the standards its federal awards attach.

Answering them requires no new rule. They require the program office, the contracting office, the security team and the budget office to agree on who owns each job. In our view, that agreement is what the 2018 strategy asked for, even though it was written as a security document.

There is also a sequencing point. The four jobs depend on each other in one direction. Exclusion is only as good as the information behind it, which the access job supplies. Access is cheaper to negotiate when the clause is uniform, because the supplier has seen it before. Uniform clauses are easier to defend when the government meets the same standard itself. An agency that starts with example, by fixing its own systems first, makes each of the other three jobs easier to argue for.

Finally, each job needs a measure that a leader can review. For exclusion, the measure is the time from a risk decision to the last purchase of the excluded item across all vehicles. For access, it is the share of critical supplier systems where the contract grants the access the security team would need during an incident. For uniformity, it is the number of security clause variants in use and the number with a recorded reason. For the example job, it is the list of standards the agency demands of others and does not yet meet itself. None of those measures appears in the strategy. They are our suggestion for turning its four promises into something a program office can track from one quarter to the next.

With hindsight

The exclusion job got its statute that December. On December 21, 2018, the SECURE Technology Act became Public Law 115-390. Its title II, the Federal Acquisition Supply Chain Security Act, amended title 41 of the United States Code to provide for federal acquisition supply chain security. It established a Federal Acquisition Security Council and required it to set criteria and procedures for recommending exclusion orders, which exclude sources or covered articles from executive agency procurement actions, and removal orders, which require the removal of covered articles from executive agency information systems.

The governmentwide contract clause took longer. On October 5, 2023, the Defense Department, the General Services Administration and NASA published an interim rule amending the FAR to implement supply chain risk information sharing and exclusion or removal orders under the 2018 act, effective December 4, 2023. Under it, an offeror represents, after a reasonable inquiry, that it does not propose to provide or use any covered article subject to an order. Orders are searchable in the System for Award Management. Contractors must report to the contracting officer once they become aware that a covered article subject to an order has been delivered or used, including under orders issued after award. For existing indefinite-delivery contracts, contracting officers were to add the new prohibition clause within 6 months of December 4, 2023, to apply to future orders. The rule also changed the time for the initial report under an existing reporting clause from one business day to three business days to align with the new clause.

The contractor job moved through defense contracting. In its 2024 program rule, the Defense Department said that it had relied on offeror representation that the requirements of NIST Special Publication 800-171 had been met, and that in 2019 it announced the development of the Cybersecurity Maturity Model Certification (CMMC) to move away from a “self-attestation” model. The CMMC program rule, published October 15, 2024 and effective December 16, 2024, established three levels to verify that contractors had implemented required measures to protect federal contract information and controlled unclassified information. The DFARS rule that put CMMC into contracts was published September 10, 2025 and took effect November 10, 2025. It phases in: until three years after its effective date, the clause applies when program managers decide to apply a CMMC requirement; after that, it applies whenever a contractor will use its own information systems to process, store or transmit such information. The program those rules describe verifies that required measures are in place, through self-assessments, certification assessments and affirmations. It sets three levels: basic safeguarding of federal contract information at Level 1, broad protection of controlled unclassified information at Level 2, and higher-level protection against advanced persistent threats at Level 3. That answers the first half of the strategy’s contractor promise, reviewing contractor practices. Testing, hunting and sensoring on contractor systems, the second half, is a different and more intrusive activity. The phase-in has also slowed. A Defense Department class deviation, 2026-O0025 Revision 3, tells contracting officers to remove or revise CMMC requirements in line with a Department of War Chief Information Officer memorandum dated July 13, 2026. As the deviation describes it, that memorandum suspends the November 2026 CMMC Phase 2 transition, permits Level 1 (Self) or Level 2 (Self) assessments in all procurement requests and requires baseline compliance with NIST SP 800-171 Revision 2 under DFARS 252.204-7012.

For an agency leader in 2026, the pattern holds. The jobs with a statute or a regulation behind them moved: exclusion through the 2018 act and the 2023 FAR rule, and contractor assessment through the CMMC rules, though more slowly than those rules first set out. The jobs that depended on each agency’s own contracts and budgets, operational access to supplier systems and meeting the government’s own standards, moved as fast as each agency did. A strategy can name a job; only an instrument makes someone do it.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • National Cyber Strategy
  • supply chain risk
  • federal contractors
  • FASCSA
  • CMMC
  • procurement