Texas’s ransomware day made the remote admin path an agency risk

When one actor hit more than 20 Texas entities at once in August 2019, the state’s follow-up advice was about remote administration. Whoever operates that path, the small agency still owns its risk.

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

When one actor hit more than 20 Texas entities at once in August 2019, the state’s follow-up advice was about remote administration. Whoever operates that path, the small agency still owns its risk.

  1. The week On the morning of August 16, 2019, more than 20 entities in Texas reported a ransomware attack.
  2. Who holds the remote administration path An agency can produce a list of its IT vendors from its accounts payable records.
  3. How the path admits people The first of DIR’s incident-specific recommendations was to allow authentication to remote access software only from inside the provider’s network.
  4. What the path can do once someone is in A remote management platform exists to do things to many machines at once: install software, run scripts, change settings.
  5. Recovering when the path is turned against you When the administration path is used to deploy ransomware, it reaches whatever that path can reach.
  6. Where the argument is weakest DIR did not publish a finding on how the attackers got in.
  7. How it plays out by kind of organization For the smallest governments, the provider may be the whole IT function.
  8. With hindsight Federal attribution came on November 8, 2021.

The week

On the morning of August 16, 2019, more than 20 entities in Texas reported a ransomware attack. According to the Texas Department of Information Resources (DIR), the state’s technology agency, most of them were smaller local governments. Later that morning the State Operations Center was activated. DIR’s September update said it went to Level II, Escalated, and that DIR led the state’s response.

DIR published an update on the incident as of about 5:00 p.m. central time on August 17. It said the evidence gathered so far pointed to a single threat actor. Twenty-three entities had been confirmed as affected, and responders were working with them to bring systems back online. The State of Texas’s own systems and networks had not been affected. DIR said it appeared that every entity actually or potentially affected had been identified and notified.

The list of responders was long. It named DIR, the Texas Division of Emergency Management, the Texas Military Department, the Texas A&M University System’s Security Operations Center and Critical Incident Response Team, and the Department of Public Safety’s cyber and intelligence units. It also named the Public Utility Commission, the Department of Homeland Security, the FBI and the Federal Emergency Management Agency. The update ended with general advice for staff: be wary of unexpected links and attachments, check sender addresses, use strong unique passwords, report anything suspicious and take awareness training.

DIR posted a fuller account on September 5. By 7:00 p.m. on August 23, it said, every affected entity had moved from assessment and response to remediation and recovery, with business-critical services restored. That transition ended the State Operations Center’s role. More than half of the entities were back to normal operations, and DIR was scheduling follow-up visits to check their rebuilds. DIR said it was unaware of any ransom being paid.

DIR described the event as the first coordinated state and federal response of its kind to a statewide, multi-jurisdictional cyber incident. A response plan prepared by its Office of the Chief Information Security Officer had been ready. Teams were in the field at the worst-hit sites within hours of notice. By day four they had visited every affected site and finished state response work at more than a quarter of them. One week after the attack began, all sites were cleared for remediation and recovery.

The September update also changed the advice. The general hygiene list of August 17 gave way to specific steps for this incident, addressed to anyone whose servers or computers are administered remotely, by internal IT staff or by a managed service provider. Allow authentication to remote access software only from inside the provider’s network. Use two-factor authentication on remote administration tools and on virtual private network tunnels, rather than remote desktop protocol. Block inbound traffic from Tor exit nodes and outbound traffic to Pastebin. Use endpoint detection and response to catch PowerShell running unusual processes.

Neither DIR update named the entities or the ransomware, and the September update said the federal investigation was continuing. Federal charges came in November 2021, and are covered in the hindsight section below.

The early public numbers out of Texas were counts of victims. Twenty-three entities confirmed by the afternoon of August 17, most of them small local governments, all hit by one actor at once. The count was alarming, but it was the least useful fact for anyone deciding what to change. A town hall, a county office and a small utility do not usually fall on the same morning by coincidence. When they do, something they share has been used against them.

The state’s own account pointed to what that shared thing was likely to be. Its August 17 advice was generic: watch for phishing, use strong passwords, take training. Its later advice, framed as specific to this incident, was about remote administration. It told anyone whose systems are run remotely, by in-house staff or a managed service provider, to restrict where remote tools accept logins from, to require a second factor, and to watch for the scripts an intruder runs once inside.

Our argument is that the Texas attack made the remote administration path a risk an agency has to own, whoever operates it. A small agency may not run its own IT at all. Somebody, somewhere, holds a tool that can reach every server and laptop it has, and push software to them. That path is the most powerful account in the building. It can be the least examined, when it belongs to a contractor, or to a part-time employee, or to a product nobody remembers buying.

We break the issue into four decisions about that path, which do not overlap and between them cover it. First, who holds it: the people, firms and products with remote administrative reach. Second, how it admits people: where logins may come from and what proves identity. Third, what it can do once someone is in: the commands it can run and how fast misuse is noticed. Fourth, how the agency recovers when the path is turned against it. A fifth section tests the argument against the strongest objections. The last looks at how it plays out for different kinds of organization.

The sector here is government because the victims were mostly local governments, and because public bodies of that size can depend heavily on outside administration. Banks, credit unions and health systems that use managed providers face the same structure, and the last section turns to them. Contract and regulatory questions differ by jurisdiction, so agencies should check specifics with counsel and their state’s IT authority.

Who holds the remote administration path

An inventory of paths, not of vendors

An agency can produce a list of its IT vendors from its accounts payable records. A list of every route by which someone outside the building can run commands on agency machines is harder to build. The second list is the one that matters here, and it can be longer than the first.

It includes the managed service provider’s remote monitoring and management tool, installed as an agent on every endpoint. It includes the remote desktop gateway a former contractor set up and nobody switched off. It includes support tools that a line-of-business vendor, such as a utility billing or court records provider, uses to troubleshoot its own application. It includes the virtual private network that staff use from home, and the accounts on it. Each one is a door, and each one has a key holder.

Building the inventory is unglamorous work. An agency can start from its endpoints: list the remote access and management agents installed on each machine. It can then work from its firewall: list the inbound connections it permits and the services that answer them. Finally, it can work from its contracts and invoices: list every party paid to support a system remotely. Where the three lists disagree, there is a door nobody is accounting for.

What the contract should say

For a small agency that outsources, the managed provider may be the IT department. The contract is then the main instrument of security policy. A contract written around help desk response times and patching cycles can be silent on how the provider protects its own access to the customer.

A contract that treats the administration path as a shared risk would cover a short list of points. How the provider’s staff authenticate to the remote tools, and whether a second factor is required. Whether the provider’s tool accepts logins only from the provider’s own network. Whether administrative credentials are unique to each customer. How quickly the provider must tell the agency about a compromise of its own systems. What logs the agency can obtain, and for how long they are kept. Who pays for and leads incident response if the provider’s tool is the route in.

These clauses do not require an agency to be technically sophisticated. They require it to ask, in writing, and to keep the answers. A provider that cannot answer them is telling the agency something important.

Shared providers concentrate exposure

The same arrangement that makes outsourcing affordable for small governments also concentrates their risk. A provider that serves many towns in a region can run one console that reaches all of them. A compromise of that console is a compromise of every customer it can reach, on the same morning.

That is the mechanism that turns one intrusion into a multi-jurisdiction event, and it fits the pattern Texas reported: one actor, many small entities, all at once. It also means an agency cannot judge its exposure by looking only at itself. It has to ask how many other customers share its provider’s tooling, and what separates them. The answer affects how an agency weighs a cheaper shared service against a more isolated one.

Ownership of this decision should sit with a named person. In a county that may be the IT director. In a town of a few thousand people it may be the city manager or the clerk, with the provider’s account manager as the counterpart. What matters is that someone inside the agency is accountable for knowing who holds the path, rather than assuming the provider has it covered.

Products that bring their own remote access

Some remote paths arrive without anyone deciding to create them. A line-of-business application may be sold with a support agreement that assumes the vendor can connect whenever a ticket is opened. A building system, a camera platform or a phone system may include a cloud management portal that reaches into the network. Each was bought for its function, and its remote access came along as a feature.

These paths deserve the same scrutiny as the managed provider’s console, scaled to their reach. A vendor tool that can only touch one application server is a smaller risk than one that can reach the domain. The procurement step is the cheapest place to catch them. A short question in every purchase of software or connected equipment does the job: will the vendor need remote access, how will it authenticate, and can the agency switch it off between support sessions.

Measures that show the path is under control

A small agency does not need a dashboard, but it does need a few numbers it can check each quarter. The count of remote administration routes in the inventory, and how many were added or closed since the last review. The share of those routes protected by a second factor. The number of exempt accounts and the age of each exemption. The time taken to remove access when a provider technician or employee leaves.

One more measure tests the relationship rather than the technology: how long the provider takes to answer a written security question. A provider that answers in days, with evidence, is behaving like a partner in the risk. A provider that does not answer is a finding in its own right, and belongs in the agency’s risk register with a named owner.

How the path admits people

Restricting where logins come from

The first of DIR’s incident-specific recommendations was to allow authentication to remote access software only from inside the provider’s network. The logic is simple. A remote management console that accepts logins from anywhere on the internet can be attacked from anywhere. One that accepts them only from a known network forces an attacker to get into that network first.

For an agency, the practical question is whether its provider has done this, and how the agency would know. A provider can describe its configuration. Better, the agency can ask for evidence: the allow-list in force, or a statement of the network ranges the tool accepts, refreshed when it changes. The same logic applies to any vendor support tool. A billing vendor that connects from its own offices can be restricted to those addresses.

The restriction has costs. Provider staff who work from home or on the road need a route back into the provider’s network before they can reach customers. That adds a step during an outage, exactly when speed matters. The trade is usually worth it, but it should be made deliberately, with an emergency path that is itself protected.

A second factor on the tools that matter most

DIR’s second recommendation was two-factor authentication on remote administration tools and virtual private network tunnels, rather than remote desktop protocol. Remote desktop exposed to the internet relies on a password alone unless something else is added. Passwords leak, get reused and get guessed.

The order of rollout matters. An agency with limited staff should start where the reach is greatest: the provider’s console and any account that can push to every machine. Next come remote access accounts for staff and vendors. Email and ordinary user accounts follow. Starting with email because the feature is easy to switch on there gets the order backward.

Two details can undo the control. Shared accounts, used by several provider technicians, make a second factor meaningless if the factor is shared too. Exemptions granted for convenience, such as a service account or a legacy tool that cannot take a second factor, become the path of least resistance. An agency should ask its provider to list any exempt accounts and the reason for each.

Emergency access without a back door

Every restriction above raises the same practical worry. What happens when the provider’s network is down, the second factor device is lost, or the one technician who knows the system is unavailable during an outage. The temptation is to keep a spare route open for emergencies, with a shared password and no second factor.

A better pattern is a break-glass account that is protected differently rather than less. Its credentials are sealed and stored where only named people inside the agency can reach them. Its use triggers an alert to the agency and the provider. Its password is changed after every use. The agency gets a dependable emergency route without leaving an unguarded door for the rest of the year.

Closing the doors nobody uses

DIR also advised blocking inbound traffic from Tor exit nodes. Tor is an anonymizing network, and a small local government is unlikely to need administrative connections from it. Blocking it does not stop a determined attacker. It removes a cheap form of cover.

The broader version of this step is to close remote access routes that are no longer needed. The remote desktop gateway left by a past project, the vendor support tool for a system since retired, the staff remote access account of someone who left. Each is cheap to close and expensive to leave open. A quarterly review against the inventory described above is, in our judgment, a reasonable cadence for a small agency.

Proving it to the agency

Each of these controls is invisible from the agency’s side unless the provider shows it. An agency should not have to take the provider’s word, and it may not need a full audit either. A middle course can work: an annual written attestation from the provider covering the controls in the contract, a screenshot or export of the relevant settings, and the right to ask for more after an incident in the provider’s customer base.

Where a provider holds an independent assurance report, the agency should read the parts that cover remote access to customers, rather than stopping at the cover letter. Such reports describe what was tested and what was excluded. If the customer-facing management platform was out of scope, the report says little about the path that matters here.

What the path can do once someone is in

Remote tools are built to run commands at scale

A remote management platform exists to do things to many machines at once: install software, run scripts, change settings. Those are exactly the capabilities an attacker needs to deploy ransomware across an estate. An intruder who controls the console does not need to break into each machine. The console does it on the attacker’s behalf, using permissions the agency granted.

That changes what “least privilege” means. It is not enough that the provider’s technicians have only the access they need. The tool itself should be configured to do only what the agency needs. If the provider never pushes arbitrary scripts to a customer’s servers, the feature can often be restricted or require a second approval. If mass deployment is needed for patching, it can be limited to approved packages.

Watching for the scripts attackers run

DIR’s last two recommendations dealt with what happens after entry. Block outbound traffic to Pastebin, a public text-sharing site. Use endpoint detection and response tools to detect PowerShell running unusual processes. PowerShell is the scripting environment built into Windows. Administrators rely on it, and so do intruders.

The point of both controls is time. Ransomware deployed through a management tool can encrypt many machines quickly. The earlier unusual activity is seen, the more of the estate can be saved. For a small agency with no one on staff to watch alerts at three in the morning, that may mean buying detection as a service.

Buying it creates its own dependency. The detection provider may be the same firm as the management provider. If one firm both administers the systems and watches for misuse of that administration, a compromise of the firm may blind both functions at once. Separating them costs more. Where an agency cannot afford separation, it should at least ask whether the detection console uses different credentials and infrastructure from the management console.

The agency’s own accounts matter too

Attention on the provider should not hide the agency’s own administrative accounts. Staff who hold local administrator rights on their workstations, shared accounts used by a department, and service accounts created for old projects all widen what an intruder can do once inside. A ransomware operator who reaches one machine looks for exactly these.

The fixes are familiar and cheap. Remove local administrator rights from everyday accounts. Give the few people who need administrative power a separate account for that purpose. Retire service accounts that no system still uses. Each step shrinks what the administration path, or anyone who steals part of it, can reach.

Keeping one customer’s breach from becoming another’s

A provider that serves many agencies should separate them. Each customer’s environment should be reachable only with credentials specific to that customer, so that a password stolen in one town does not open the next. Administrative accounts should not be shared across customers, and the provider’s own corporate network should not double as the route into every client.

The agency cannot build this separation itself, but it can ask how it is done and whether a single technician account can reach every customer at once. The answer tells the agency how far a compromise elsewhere would travel. A provider with strong separation turns a regional event into a contained one. A provider without it puts every customer on the same morning.

Evidence the agency should be able to see

An agency that holds the risk should be able to see the record. That means logs of who connected through the management tool, when, from where, and what they ran. Those logs may live in the provider’s system, not the agency’s. If the agency cannot obtain them on request, it cannot check the provider’s account of an incident, and investigators may not get them in time.

A modest standard is enough to start. Logs of remote administrative sessions kept for a stated period, exportable by the agency, and preserved automatically when an incident is declared. The contract should say so. The agency should test the request once, in calm conditions, so the process is known before it is needed.

Recovering when the path is turned against you

Recovery depends on what the attacker could not reach

When the administration path is used to deploy ransomware, it reaches whatever that path can reach. Backups that are managed through the same console, stored on the same network or reachable with the same credentials are at risk with everything else. The recovery question is therefore whether any copy of the data sits outside the path the attacker used, as well as whether backups exist at all.

For a small agency, we would want at least one copy that is offline or held by a separate service, under separate credentials. It also means knowing how to rebuild critical systems from scratch: which machines matter first, where installation media and licenses are, and who can do the work if the regular provider is itself compromised. Restoring a utility billing system or a records system is slow if the documentation lived on the encrypted servers.

Testing the rebuild before it is needed

A backup that has never been restored is a hope rather than a control. A small agency can test restoration on a modest schedule: pick one important system each quarter, restore it to a separate environment, and record how long it took and what was missing. The record becomes evidence for auditors and insurers, and the gaps it reveals become the next quarter’s work.

The test should include the case where the regular provider is unavailable. Someone other than the provider should know where backups are held, how to reach them and which credentials unlock them. Those credentials should be stored somewhere the provider’s administration tool cannot reach.

Keeping services running for residents

Ransomware in a local government stops services residents notice: payments, permits, court dockets, utility billing. DIR measured the Texas recovery partly by when business-critical services were restored. An agency’s plan should list those services, the manual workaround for each, and how residents will be told what is available. Paper forms, extended deadlines and a public notice on a site hosted outside the affected network are cheap to prepare and hard to improvise.

What Texas had in place

DIR’s account of the response is instructive for what it shows about preparation. The state had a response plan before the event. It activated the State Operations Center the same morning and had state and federal teams in the field at the worst-hit sites within hours of notice. Its teams had visited every affected site by day four. It published a confirmed count of affected entities on August 17.

That capacity belonged to the state, not to the towns. The affected entities drew on a state agency, a university security operations center, the Texas Military Department and federal partners. A town’s recovery plan in Texas, in effect, included the state’s plan. That is a sound arrangement, provided the town knows how to invoke it.

Knowing whom to call, and in what order

An agency’s incident plan should name the calls it will make in the first hour: its provider, its cyber insurer if it has one, the state IT or emergency management authority, and law enforcement. It should say who is allowed to make them. It should also say what to do if the provider is unreachable or compromised, because that is the likely case when the provider’s tool was the route in.

Preserving evidence belongs in the plan too. DIR’s leadership credited the affected entities with resolving the incident quickly while protecting the integrity of the federal investigation. Rebuilding fast and preserving evidence pull in different directions. An agency that wipes and restores everything at once may lose what investigators need. The plan should say who decides the balance, and the state or federal responders can advise.

Rebuilding the path as well as the machines

DIR’s September update said it was scheduling follow-up visits to make sure the affected entities’ rebuilds had worked. That step deserves emphasis. Restoring servers and laptops from backup does not repair the route the attacker used. If the same remote access path, with the same accounts and the same settings, comes back with the restored systems, the agency has rebuilt its exposure along with its services.

A sound rebuild resets every credential that touched the administration path, reissues the provider’s access under the tightened rules, and confirms that unused routes stay closed. It also reviews the logs from the incident to see what the attacker did and where else it might have reached. The follow-up review is the moment to make the changes in this article, because the cost of not making them has just been demonstrated.

Paying, or not

DIR said it was unaware of any ransom paid in the Texas event. Whether to pay is a decision for each victim, with counsel, insurers and law enforcement involved, and with attention to sanctions rules. What preparation changes is whether the question arises at all. An agency with clean, reachable backups and a practiced rebuild plan has a real choice. One without them may have none.

Where the argument is weakest

The vector was not formally published

DIR did not publish a finding on how the attackers got in. Its incident-specific advice addressed remote administration, and that is strong circumstantial evidence. It is not a forensic report. A skeptic can fairly say that the argument rests on an inference from the state’s recommendations.

The answer is that the argument does not depend on the exact vector in Texas. The remote administration path is the most powerful access route in a small agency whether or not it was used that day. The four decisions above are worth making on that ground alone. Texas made the risk visible; it did not create it.

Small agencies cannot do all of this

A town with no IT staff cannot build inventories, negotiate security clauses and buy separate detection. That objection is largely right. A small agency may have a limited budget and limited expertise, and some of the steps above require both.

Two things soften it. First, most of the steps are questions to ask a provider, not systems to build. Asking for second factors on the provider’s console, for logs, and for an exempt-accounts list costs little. Second, the burden can be pooled. State IT agencies, councils of government and cooperative purchasing programs can write model contract terms once and let small agencies adopt them.

Providers will resist

Some managed providers will resist security terms, or price them high. A small agency may have few alternatives in its region.

Resistance is itself information. A provider unwilling to describe how it protects its own access is asking the customer to accept an unknown risk. Where alternatives are scarce, collective purchasing helps again. A provider that would refuse one town’s clause may accept it from a state contract that covers dozens of customers.

The real risk is phishing

DIR’s August 17 advice in the Texas event was about email: suspicious links, attachments and impersonation. Phishing was the threat that advice addressed, and a skeptic could argue that attention belongs there rather than on remote administration.

The two are not alternatives. A phishing email may be how an attacker first gets into a provider or an agency. The remote administration path is what lets that first foothold reach every machine at once. Training reduces the chance of the first step; controls on the path limit the damage when the first step succeeds anyway. An agency with little money should fund both, and the path controls are cheaper per machine protected.

Outsourcing was supposed to reduce risk

The case for a managed provider is that it can do security better than a part-time administrator. That case can be sound. A good provider patches faster, monitors more closely and recovers better than a single small agency could.

The argument here does not say that outsourcing is a mistake. It says the risk does not leave with the work. A provider concentrates both competence and exposure. The agency gains the competence and still holds the exposure, so it has to understand it.

How it plays out by kind of organization

Towns and small cities

For the smallest governments, the provider may be the whole IT function. The priority is the contract and a short list of questions: who can reach our machines, how they authenticate, what logs we can see, and what happens if your systems are compromised. A clerk or manager can ask those questions without technical training. The state IT authority’s incident contacts belong in the same folder as the contract.

Small governments also gain from comparing notes. Neighboring towns that use the same provider can ask their questions together, compare answers and share the cost of a review. If the provider’s console is a shared exposure, shared scrutiny is the proportionate response.

Counties and larger cities

Larger local governments may have internal IT staff as well as vendors. Their path inventory is longer, because departments buy their own systems with their own support arrangements. Sheriff’s offices, courts and public health departments may each have a vendor with remote access. Central IT should hold the inventory and approve new remote access routes, even when a department pays the bill.

Special districts and utilities

Water districts, school districts and small utilities may run operational technology alongside office systems. Remote access for an operational vendor can reach equipment that affects public services. These entities should treat any remote path into operational systems as separate from the office network, with its own approval and logging. They should check sector-specific requirements with their regulators.

Governments that act as providers

Some counties and regional councils provide IT services to the smaller towns and districts around them. In that role they are the managed provider, and everything in this article applies to them from the other side. They hold the console that reaches many jurisdictions. They should apply to themselves the separation, authentication and logging they would demand of a private firm, and they should write those commitments into their interlocal agreements.

State agencies

State agencies are on both sides. They hold their own remote administration paths and must manage them. They also act, as DIR did, as the responder of last resort for local governments. That second role argues for state model contract terms, shared detection services and tested mutual aid plans, so that the capacity Texas showed is planned rather than improvised.

Banks, credit unions and health systems

Regulated private organizations face the same structure. Community banks, credit unions and ambulatory practices may outsource IT administration too. Where their regulators expect oversight of third-party providers, the remote administration path is one of the most important things that oversight should cover. The questions in the first section translate directly into vendor due diligence, with the contract and the provider’s evidence as the record an examiner or auditor would ask to see.

One difference is worth drawing out. A bank or health system that suffers ransomware through its provider faces customer or patient notification duties, depending on the data involved and the laws that apply. A local government faces public records and transparency expectations alongside any notification law. In both cases the organization, not the provider, answers to the public. That is the strongest reason the administration path belongs on the organization’s own risk register.

With hindsight

Federal attribution came on November 8, 2021. The Justice Department announced charges against a Russian national, alleging he conducted Sodinokibi/REvil ransomware attacks against multiple victims, including businesses and government entities in Texas on or about August 16, 2019. It also announced the seizure of $6.1 million in funds it said were traceable to ransom payments he had received. He was believed to be abroad. The charges included conspiracy to commit fraud and related activity in connection with computers, damage to protected computers and conspiracy to commit money laundering. The seizure warrant was issued in the Northern District of Texas, and the FBI’s Dallas and Jackson field offices led the investigation. The release said that, according to the indictments, the defendants had accessed victim companies’ internal networks and deployed the ransomware to encrypt their data.

The same announcement covered a second defendant, a Ukrainian national arrested in Poland and charged over the July 2021 attack on Kaseya, an IT software company. According to the department, that attack pushed REvil code through a Kaseya product so that the product deployed ransomware to endpoints on its customers’ networks. That was the remote administration risk at international scale. A tool trusted to manage many machines became the means of encrypting them.

Guidance caught up. In May 2022, cybersecurity agencies of the United Kingdom, Australia, Canada, New Zealand and the United States, including CISA, the NSA and the FBI, issued a joint advisory on threats to managed service providers and their customers. It said attackers trying to compromise providers exploit vulnerable devices and internet-facing services, run brute force attacks and use phishing. It told customers to enforce multifactor authentication on provider accounts that access their environment, and to remove accounts no longer in use. It said contracts should clearly define ownership of security roles and responsibilities. It recommended segregating customer environments, not reusing administrative credentials across customers, keeping important logs for at least six months, and storing backups isolated from the network.

Money followed for local governments. On September 16, 2022, the Department of Homeland Security announced the State and Local Cybersecurity Grant Program. With a companion tribal program, it was set to make $1 billion available over four years. CISA describes the program as designed to direct funding to the least mature entities, local governments, with states passing at least 80 percent of their funding to them. DHS announced $91.7 million for the fiscal year 2025 round on August 1, 2025.

Texas reorganized. The 89th Texas Legislature created the Texas Cyber Command through House Bill 150 and moved cybersecurity functions to it from DIR. In an April 9, 2026 notice, DIR said the command had become the state’s centralized authority for cybersecurity and that the transition would proceed in phases. It told state agencies and local governments that services and reporting processes were unchanged at that point.

What held up is the shape of the risk. The 2022 joint advisory reads as a fuller version of DIR’s September 2019 list, and the Kaseya case showed the path working exactly as the Texas advice implied it could. What remains open is the small agency’s side of the bargain. Grants and state commands add capacity, but neither writes an agency’s contract or inventories its remote access for it. Those decisions still sit with each agency.

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

  • ransomware
  • local government
  • managed service providers
  • remote access
  • incident response