Rule 41’s 2016 change let one judge approve remote access to hacked victims’ computers, with notice afterward
From December 1, 2016, a magistrate judge could issue one warrant for remote access to damaged computers in five or more districts, with reasonable efforts at notice, electronic if need be. Later court-authorized FBI operations removed malware from private computers and sought to notify owners.
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
From December 1, 2016, a magistrate judge could issue one warrant for remote access to damaged computers in five or more districts, with reasonable efforts at notice, electronic if need be. Later court-authorized FBI operations removed malware from private computers and sought to notify owners.
- The week The Chief Justice transmitted the amendments to Congress the same day, under the Rules Enabling Act at section 2072 of title 28. [1] The transmittal listed the materials that came with them.
- Venue: where a warrant may be issued, and what did not change Before the amendment, the department said, the rules generally required a search warrant application in the district where the property was located. [2] That is easy to apply to a house.
- The victim’s machine: why paragraph (b)(6)(B) reaches hacked computers Subparagraph (A) is about suspects who hide.
- Notice: what the rule requires the government to tell the owner Rule 41(f)(1)(C) already required the executing officer to give a copy of the warrant and a receipt to the person from whom, or from whose premises, property was taken. [1] Or the officer could leave them at the place where the property was taken. [1] That works for a physical search.
- What stays with the organization An operation under a remote access warrant acts on the government’s investigative purpose.
- How the amendment lands on different organizations For the health system, a remote operation could touch a server that holds protected health information.
- Objections, in order The objection is that letting a single magistrate judge authorize access to computers across the country removes the check that local judges provided.
- Checks a CISO or general counsel could run These checks follow from the rule and the department’s descriptions.
- With hindsight Later operations showed what court-authorized remote action against victims’ systems looked like.
The week
On December 1, 2016, amendments to Rule 41 of the Federal Rules of Criminal Procedure took effect. [1] The Supreme Court had adopted them by order on April 28, 2016, together with amendments to Rules 4 and 45. [1] The order said the amendments would govern criminal proceedings commenced after that date and, insofar as just and practicable, proceedings then pending. [1]
The Chief Justice transmitted the amendments to Congress the same day, under the Rules Enabling Act at section 2072 of title 28. [1] The transmittal listed the materials that came with them. They included a transmittal letter to the Court dated October 9, 2015, a redline version with Committee Notes and excerpts from committee reports of May and September 2015. [1]
The Rule 41 changes had two parts. The first added a new paragraph (b)(6). [1] It let a magistrate judge with authority in any district where activities related to a crime may have occurred issue a warrant to use remote access. [1] The warrant could authorize searching electronic storage media and seizing or copying electronically stored information inside or outside that district. [1]
The new paragraph applied in two situations. One was where the district holding the media or information had been concealed through technological means. [1] The other was an investigation of a violation of 18 U.S.C. 1030(a)(5), where the media were protected computers damaged without authorization and located in five or more districts. [1]
The second part changed notice. For a remote access warrant, the officer had to make reasonable efforts to serve a copy of the warrant and receipt. [1] The person to be served was the one whose property was searched or who possessed the information seized or copied. [1] Service could be by any means, including electronic means, reasonably calculated to reach that person. [1]
In the days before the effective date, the Justice Department argued for the change in a series of blog posts by the head of its Criminal Division. On November 21, 2016, the first post addressed anonymizing technology. [2] It said a rule first codified in 1917 generally required a search warrant application in the court where the property was located. [2] It said that when a suspect’s location was hidden, the rules did not clearly identify which court investigators should use. [2]
That post described the department’s view of what was at stake. It said that in several cases arising from one investigation of a hidden child exploitation site, federal courts had thrown out evidence because agents could not guess the right court. [2] It said the amendments did not create new law enforcement authority. [2]
On November 28, 2016, the third post responded to critics. [3] It said the amendments addressed venue only, identifying which court could consider a warrant application. [3] It said they did not change constitutional requirements of probable cause and particularity, or the statutory rules for real-time interception. [3]
The same post addressed botnet victims directly. It said searches of infected computers would be done to investigate the extent of a botnet, or to obtain information needed to free victims’ computers from it. [3] It gave examples: obtaining identifying information to contact owners and warn them, or an online operation to disrupt the botnet and restore owners’ control. [3]
It also acknowledged risk. The post said law enforcement actions online could never be completely free of the risk of unintended consequences. [3] It said investigators conducted pre-deployment verification and validation of tools before complex operations. [3] It said the department welcomed continued work with the private sector on botnet mitigation. [3]
Rule changes about venue sound procedural. The December 1, 2016 change to Rule 41 was procedural on its face. [1] For a hospital, bank or agency that runs internet-facing systems, though, it changed something practical. It made a single court the place where federal investigators could seek permission to reach into many hacked machines at once, including theirs.
Our argument is that the 2016 amendment made an organization’s compromised server easier to reach by warrant, as one of many machines in a single application, and told the organization about it afterward. [1] The rule did not change what a warrant needs. [4] It changed where one could be issued and how the owner would hear about it. [1] In our reading, those two changes made the victim a likelier subject of a search, with notice that might arrive by email.
We take the amendment in four parts:
- Venue: what changed about where a warrant may be issued, and what did not change.
- The victim’s machine: why paragraph (b)(6)(B) reaches computers that were hacked, not computers of suspects.
- Notice: what the rule requires the government to tell the owner, how and when.
- What stays with the organization: the work the rule does not do, from patching to verifying a notice.
The first part is the court, the second the target, the third the communication and the fourth the remaining duty. Together they cover the amendment’s effect on an organization whose systems might be searched. After them we look at how it lands on different kinds of organizations, take the strongest objections in order, and end with checks a CISO or general counsel could run.
Venue: where a warrant may be issued, and what did not change
The old default
Before the amendment, the department said, the rules generally required a search warrant application in the district where the property was located. [2] That is easy to apply to a house. It is harder to apply to a server whose location is hidden, or to thousands of infected computers spread across the country.
The Justice Department described two problems. When a suspect hid a computer’s location with anonymizing tools, investigators could not know which court to ask. [2] For botnets, it described the need to identify a single court that could hear a warrant application for a nationwide investigation. [3]
The new paragraph
Paragraph (b)(6) addressed both. A magistrate judge with authority in any district where activities related to a crime may have occurred could issue a remote access warrant for media or information inside or outside that district. [1] Subparagraph (A) covered concealed locations. [1] Subparagraph (B) covered damaged protected computers in five or more districts, in investigations under the computer fraud statute’s damage provision. [1]
The rule’s caption changed too. The heading of Rule 41(b) went from authority to issue a warrant to venue for a warrant application. [4] The Committee Note said the change made clear that Rule 41(b) identifies which courts may consider an application, not the constitutional requirements for issuing a warrant, which must still be met. [4]
Subparagraph (A): hidden locations
The first new venue rule was aimed at suspects, not victims. The department’s November 21 post described sites on the Tor network that hid their users’ locations. [2] It described an investigation in which authorities took control of such a site and obtained court approval to use a remote search tool. [2]
According to the department, the tool caused a user’s computer to send investigators limited information, including its true IP address, if the user accessed illegal material on the site. [2] The department said several courts in resulting cases had suppressed evidence because the warrant came from a court that might not be the right one. [2] It said other courts had declined to suppress but suggested they might in future cases. [2]
Subparagraph (A) answered that problem by making any district where activities related to the crime may have occurred a proper venue when the location is concealed. [1] For most organizations, this part of the rule matters less directly. Their systems are not hiding their location. It matters if an attacker uses anonymizing tools to run infrastructure on, or through, an organization’s systems.
The process behind the rule
The amendment came through the federal judiciary’s rulemaking process. The Supreme Court’s transmittal listed the materials submitted for its consideration. [1] They were a transmittal letter dated October 9, 2015, the redline with Committee Notes, and excerpts from the advisory committee’s May 6, 2015 report and the standing committee’s September 2015 report. [1]
The department described the process as three years of public deliberation, with hearings and memorandums in which objections were raised and answered. [3] It said the committee members were mostly federal judges, with state judges, academics and defense attorneys. [3] For organizations, the point is that the rule did not appear suddenly. The objections now raised by owners of systems were part of the record before it took effect.
What the note said stayed the same
The Committee Note was careful about scope. It said the amendment did not address constitutional questions, such as how specific a warrant must be for remote searches of electronic media. [4] It left those questions to ongoing case law. [4]
The Justice Department made the same point repeatedly that week. It said the amendments did not change probable cause or particularity requirements, or the statutory framework for intercepting communications in real time. [3] It said warrants under the new venue provisions would still have to be authorized by independent federal courts and could be challenged after execution. [3]
Why venue still matters to the target
A venue rule can be neutral on standards and still matter. Before the amendment, by the department’s account, a nationwide botnet investigation had no single court clearly able to hear the application. [3] After it, one judge could consider the whole application. [4] The Committee Note said that would eliminate the burden of multiple warrants in numerous districts and let a single judge oversee the investigation. [4]
For the owner of one infected server, that means the court authorizing access to its machine may sit far away, overseeing a national operation. The owner has no role in the application. Its first contact with the process is the notice.
The definitions
Subparagraph (B) borrows its key terms from the computer fraud statute. The Committee Note said that protected computer and damage have the meanings in 18 U.S.C. 1030(e)(2) and (8). [4] Those definitions are federal law outside the rule. Organizations should ask counsel how they apply to their own systems.
In our reading, an internet-connected server at a hospital, bank or agency can fall within that definition. If so, the five-district threshold, not the type of computer, is what limits the subparagraph.
The victim’s machine: why paragraph (b)(6)(B) reaches hacked computers
Damaged computers are victims’ computers
Subparagraph (A) is about suspects who hide. Subparagraph (B) is different. It applies where the media are protected computers that have been damaged without authorization. [1] The computers it reaches are, by definition, computers someone else has attacked.
The Committee Note named the kind of crime it had in mind. Activity under section 1030(a)(5), such as creating and controlling botnets, may target many computers in several districts. [4] The machines to be searched would be the ones the botnet had captured.
What the department said the searches were for
The Justice Department acknowledged that critics objected to searches of victim computers. [3] Its response described two uses. One was investigating the extent of a botnet. The other was obtaining information needed to free victims’ computers from it. [3]
Its examples were concrete. Law enforcement could gather identifying information from bot computers to contact owners and warn them. [3] Or it could run an online operation designed to disrupt the botnet and restore full control to the owners. [3] The department said both could arguably involve a search or seizure under the Fourth Amendment. [3]
The scene of a crime
The department framed the victim’s machine as a crime scene. It said existing law expressly contemplated that it may be appropriate to search property of an innocent crime victim. [3] It said it would be strange if the law forbade searching the scene of a crime. [3]
That framing is useful for an organization’s planning. A crime scene search is done for investigative purposes, not for the owner’s benefit. An operation to disrupt a botnet may help the owner. It is still the government acting on the owner’s system under a court order the owner did not see in advance.
Five districts as the trigger
The five-district threshold is a fact about the campaign, not about any one victim. [1] An organization cannot tell from inside whether its compromise is part of a campaign that reaches five or more districts. It learns that, if at all, from public reporting or from a notice.
That has a planning consequence. Any organization hit by a widespread attack should assume its systems could fall within a remote access warrant. The assumption costs little. It prompts the organization to keep its logs and to watch for unexplained changes after a public takedown announcement.
Copying as well as searching
The rule speaks of searching electronic storage media and of seizing or copying electronically stored information. [1] A remote operation may therefore take copies of what is on the victim’s machine, within the warrant’s terms. In an investigation of malware, the obvious target is the malware itself and its configuration.
For the owner, copying raises a records question. Whatever the government copies has left the organization’s control under legal process. The receipt the rule requires should describe what was taken. [1] The organization should keep the receipt with its incident records and ask counsel whether the copying affects any of its own obligations about the data involved.
Limits the department described
The department described the venue rule as narrow. It said the amendments did not authorize any search or seizure, or any remote search technique, not already permitted under the Fourth Amendment. [3] It said the Constitution already forbade mass, indiscriminate rummaging through victims’ computers and would continue to do so. [3]
It also described its own care. Before complex online operations, it said, investigators conducted pre-deployment verification and validation of tools, to make sure they worked as intended. [3] It said it strove to work closely with computer security researchers inside and outside government. [3]
What the owner cannot see
For an organization, those assurances describe the department’s process, not a right the owner can check in advance. The warrant application, the tool and the testing are on the government’s side. The owner sees the effect on its system and, later, the notice.
That makes logging and change detection more important. An organization that wants to know what was done to its server needs records of what ran on it and what changed. Those records are also what its own incident response would need.
Notice: what the rule requires the government to tell the owner
The text
Rule 41(f)(1)(C) already required the executing officer to give a copy of the warrant and a receipt to the person from whom, or from whose premises, property was taken. [1] Or the officer could leave them at the place where the property was taken. [1] That works for a physical search. A remote search has no doorstep.
The amendment added a sentence for remote access warrants. The officer must make reasonable efforts to serve a copy of the warrant and receipt on the person whose property was searched or who possessed the information seized or copied. [1] Service may be by any means, including electronic means, reasonably calculated to reach that person. [1]
Reasonable efforts, not certainty
The standard is effort, not result. The Committee Note said the amendment was intended to ensure reasonable efforts to give notice of the search, seizure or copying, and a receipt for information taken. [4] The rule does not promise that every owner will receive notice.
In a botnet case, the government may know an IP address and little else. Reaching the owner may mean going through the internet service provider. The rule’s wording, any means reasonably calculated to reach the person, allows that path. [1]
Who the person is
The rule names the person to be served as the one whose property was searched or who possessed the information seized or copied. [1] For a server in an organization’s own data center, that is the organization. For a server at a hosting provider or a managed service, the answer may be less obvious.
The hosting company owns the hardware. The customer owns the data and runs the software. A managed service provider may administer the system for the customer. Any of them could receive the notice. In our reading, the organization should not assume it will be the first to hear. Its contracts should make sure it hears promptly.
Keeping a record of notices
An organization that receives one notice may receive others. A log of law enforcement notices, with dates, systems, the warrant’s issuing court and the actions taken, gives the security team a history. It also gives auditors and regulators a clear account if they ask how the organization handled outside access to its systems.
The log should sit with legal as well as security. Some notices may come with restrictions on disclosure. Legal is best placed to decide who inside the organization may see each one.
Delay
Notice can also be delayed. The Committee Note pointed out that Rule 41(f)(3) allows delayed notice only if the delay is authorized by statute. [4] It cited 18 U.S.C. 3103a, which authorizes delayed notice in limited circumstances. [4]
So an organization cannot assume notice will arrive at the time of the operation. It may arrive later, through an intermediary, or by email from an address the organization has never seen. The rule sets a duty to try, under the court’s supervision.
Notice as the first sign of compromise
For some organizations, the notice will be the first sign that they were compromised at all. A botnet operation may reach systems whose owners never saw the infection. The department said in 2021 that many Exchange owners had removed web shells while hundreds persisted on other systems. [5]
An organization that first learns of a compromise from a government notice has to start its own response from that moment. Any clocks that run from discovery under the rules that apply to it may be affected. That is a question for counsel, and it is a reason to route notices to someone who can start the response that day.
Why the channel matters
Electronic notice brings a security problem with it. An email saying a government agency has accessed your server is exactly the message an attacker might forge. The rule does not address how a recipient should verify a notice. [1]
In our reading, that gap falls to the organization. Its incident response plan should say how to confirm a law enforcement notice through a known channel, such as a field office number taken from an official source. It should say who in legal and security receives such a notice and what they do first.
What the receipt tells you
A warrant and receipt describe what was authorized and what was taken. [1] For an organization, they are evidence. They show what changed on its system and on whose authority. They belong in the incident file next to the organization’s own logs.
They do not tell the organization how the attacker got in, what else the attacker did or whether other systems are affected. That is still the organization’s investigation.
What stays with the organization
Remediation
An operation under a remote access warrant acts on the government’s investigative purpose. The department described searches to investigate a botnet or free victims’ computers from it. [3] Neither purpose includes fixing the vulnerability that let the attacker in.
So the organization keeps the duty to patch, rebuild and verify. If an operation removes one piece of malware, the original weakness and any other tools the attacker left can remain. The organization’s own hardening work is unchanged by the warrant.
Evidence and records
A regulated organization may have its own reporting or breach analysis duties after an intrusion, depending on the rules that apply to it. In our reading, a government operation on its system does not discharge those duties. Organizations should ask counsel how their own rules treat such an event.
It may affect them. If the government copies data from the system, that copying is part of what happened to the data. The organization’s incident record should show it, with the warrant and receipt, so later reviewers can see the full sequence.
Preserving what the operation changed
A government operation changes the system it touches. Removing a web shell or uninstalling malware deletes artifacts the organization’s own investigators might want. Before rebuilding, the organization should capture what remains: disk images, memory if available, and logs covering the period before and after the operation.
That evidence answers questions the government’s notice will not. It shows how the attacker got in, whether other tools remain and what data the attacker reached. It also shows what the government’s action did, which the organization may need to explain to its own regulators or auditors.
Working with law enforcement before a warrant
The department said in 2016 that it welcomed continued collaboration with the private sector on botnet mitigation. [3] An organization that finds attacker tools on its own systems can remove them itself, and can report the intrusion voluntarily.
Acting first has two benefits. It shortens the attacker’s access. And it keeps the organization in control of its own systems and evidence, rather than learning afterward what was done under a warrant it did not see.
Contracts with providers
Notice may arrive through an internet service provider or hosting provider. [1] The organization’s contracts and contact records decide whether that notice reaches someone who can act. A provider holding an outdated security contact will forward a notice to nobody useful.
Managed service providers raise the same question one level down. If the provider runs the server, the notice may go to the provider. The contract should require prompt forwarding of any law enforcement notice about the client’s systems, subject to any legal restriction on disclosure.
Communications
An organization that learns the government acted on its systems will face questions from staff, customers and sometimes the public. The answers depend on facts the organization may not have at first. A prepared holding statement, and a rule that legal reviews any public comment on a law enforcement operation, save time.
Board and executive reporting
Leadership will want a short answer to three points: what the government did, what the attacker did, and what the organization is doing now. The warrant and receipt answer the first. [1] Only the organization’s own investigation answers the second and third.
A useful practice is to report the government’s action as one event inside the organization’s incident, not as the incident. That keeps attention on the open questions: how the attacker got in, what else they reached and when the weakness will be closed.
Decision rights
Finally, someone has to own the decision about what to do next. A notice that the FBI removed malware from a server raises immediate questions: whether the server is still compromised, whether it should come offline and who needs to know. The incident response plan should name the person who decides, and the time within which they must act.
How the amendment lands on different organizations
A health system with on-premises email
A health system that runs its own internet-facing email servers can be caught in a campaign that reaches many organizations at once, as the 2021 Exchange exploitation later showed. [5] In our reading, if attackers install tools on its servers alongside thousands of others, the conditions of subparagraph (B) could be met: damaged computers in five or more districts. [1]
For the health system, a remote operation could touch a server that holds protected health information. The warrant governs what the government may do. The health system still owns its HIPAA analysis of the intrusion itself. Its counsel should be involved from the first notice.
A bank
A bank’s internet-facing systems are just as reachable. A bank also has a strong reason to know about any outside access to its systems, for its own risk management. Its incident response plan should treat a law enforcement notice as an incident trigger, not as closure.
A county or city government
A county may depend on outside IT providers. A remote operation could reach a county server before the county knows it was compromised. The notice, routed through a provider, could be the county’s first warning.
For a county, the practical step is a current security contact with every provider and registrar, and a known person to receive law enforcement notices. Those are cheap to set up and hard to improvise.
Shared services add a layer. A county that runs systems for its towns or school districts may receive a notice about a server that holds another body’s data. The county then has to decide whom to tell and how quickly. An agreement with each served body should settle that in advance.
A federal agency
A federal agency is also a potential victim. Its systems can be swept into a botnet like anyone’s. A law enforcement notice about an agency system should flow into the agency’s own incident process, with its general counsel involved.
A software vendor whose product is the target
Some campaigns exploit one vendor’s product across thousands of customers. The 2021 Exchange operation is an example: the department described zero-day exploitation of an on-premises email product and web shells on thousands of victim computers. [5] The vendor in such a case does not own the affected servers. Its customers do.
A vendor still has a role. In 2021 the department pointed defenders to Microsoft’s remediation guidance. [5] In 2023 it listed Microsoft’s Digital Crimes Unit among partners in victim notification and remediation. [6] For a vendor serving regulated customers, a takedown is a moment to push clear remediation guidance through its own support channels, so customers do not rely on the notice alone.
An internet or hosting provider
Providers sit in the notice path. When investigators can identify an IP address but not the owner, the rule’s reasonable efforts standard may lead them to the provider. [1] A provider needs a process to receive such requests and pass notice to customers promptly. For a regulated customer, the speed of that pass-through matters.
Objections, in order
1. One judge overseeing a national operation concentrates too much power
The objection is that letting a single magistrate judge authorize access to computers across the country removes the check that local judges provided.
The Committee Note presented the single judge as a feature: one judge could oversee the investigation instead of many warrants in many districts. [4] The constitutional requirements still apply. [4] In our reading, the concern is about the depth of review a single application receives, not about venue as such. That concern is real and is a matter for the courts that review these warrants.
There is also a practical counterweight. A single judge sees the whole operation, including its scale and the tools proposed. Many judges, each seeing a slice, might never see the whole. For a victim, oversight of the entire operation by one court may be the more protective arrangement. Which is better depends on the quality of the review, which no venue rule can guarantee.
2. Government code on private systems could cause damage
The objection is that law enforcement tools running on victims’ computers could break something.
The department did not deny the risk. It said online actions could never be completely free of the risk of unintended consequences, and described pre-deployment testing to reduce them. [3] The rule itself does not set testing standards. [1] An organization should assume a small risk of side effects and watch its systems after any notice.
3. Electronic notice is unreliable and easy to fake
The objection is that email notice may never arrive, and forged notices could become a phishing tool.
Both points have force. The rule requires reasonable efforts, not guaranteed delivery. [1] It says nothing about verification. [1] The answer sits with recipients, who need a way to confirm a notice through a trusted channel before acting on it.
4. Congress, not a rules committee, should have made this change
The objection is that a change with this reach should have been legislated.
The department argued that the rules process Congress created had been used repeatedly over the years to amend Rule 41’s venue provisions. [3] The Supreme Court adopted the amendments and transmitted them to Congress under the Rules Enabling Act. [1] They took effect on December 1, 2016 and remain in the rules. [1] [8] The process question is political. The legal effect is settled by the rule now in force.
5. Victims never consented to a search of their systems
The objection is that an organization whose server was hacked should decide whether the government touches it.
The rule does not require consent. In our reading, a warrant is the legal alternative to consent. An organization that wants a say can act first: find and remove the attacker’s tools itself, or work with law enforcement voluntarily. The amendment makes delay more consequential, because a court can authorize action without the owner.
6. Operations like these leave owners complacent
The objection is that if the government cleans up botnets, owners will patch less and wait to be rescued.
The department’s own descriptions cut against complacency. The 2021 announcement said the operation did not patch the Exchange vulnerabilities or remove other tools, and urged defenders to follow remediation guidance. [5] The 2023 announcement said the action did not remediate other malware already on victim computers. [6] The 2025 announcement urged anti-virus software and security updates to prevent reinfection. [7] An owner who treats a takedown as a fix is relying on something the government did not claim to do.
Checks a CISO or general counsel could run
These checks follow from the rule and the department’s descriptions. They are our suggestions, not legal requirements.
- Name the person who receives any law enforcement notice about your systems, and publish a route for it to reach them.
- Write a verification step for such notices, using a contact taken from an official source, not from the notice.
- Confirm every internet and hosting provider, registrar and managed service provider holds a current security contact for you.
- Require providers by contract to forward law enforcement notices about your systems promptly, subject to legal limits.
- Treat a notice of a government operation as an incident trigger, with full investigation and remediation, not as closure.
- Keep logs and file integrity records on internet-facing servers that would show what changed and when.
- Add a law enforcement notice scenario to your next incident response exercise, with legal and communications at the table.
The first four make sure the notice reaches the right person. The last three make sure the organization does its own work when it does.
A tabletop scenario
The checks are easiest to test in an exercise. A simple scenario works. A security contact receives an email that appears to come from the FBI. It says a court-authorized operation removed malware from one of the organization’s internet-facing servers, and attaches a warrant and a receipt.
The exercise then asks the team to act. Someone has to verify the notice through an independent channel. Someone has to decide whether to isolate the server and preserve evidence before anything else changes. Legal has to read the warrant and receipt and decide what they mean for the organization’s own obligations.
A second injection adds a complication. The notice arrived through the hosting provider, well after the operation. Now the team has to reconstruct what happened in between, using its own logs. If the logs do not cover the period, the exercise has found its first finding.
The scenario is short. It tests contact routes, verification, evidence handling and decision rights at once. In our reading, it is a cheap way to find out whether an organization could respond to the kind of notice the 2016 rule made possible. [1]
With hindsight
The amendment took effect on December 1, 2016, as the Supreme Court’s order provided. [1] As checked on November 16, 2026, the Federal Rules of Criminal Procedure as amended to December 1, 2025 still contained Rule 41(b)(6) in the same words. [8] They still contained the 2016 sentence on reasonable efforts to serve remote access warrants, by electronic means if need be. [8]
Later operations showed what court-authorized remote action against victims’ systems looked like. On April 13, 2021, the Justice Department announced a court-authorized operation to copy and remove malicious web shells from hundreds of vulnerable computers in the United States. [5] The computers ran on-premises versions of Microsoft Exchange Server. [5]
The department said hacking groups had exploited zero-day vulnerabilities in Exchange in January and February 2021 to place web shells. [5] It said many owners had removed them, but hundreds persisted. [5] The FBI removed one early group’s remaining web shells by issuing a command through each web shell that was designed to delete only that web shell. [5]
The department was explicit about the limits. It said the operation did not patch any Exchange zero-day vulnerabilities, or search for or remove other malware or hacking tools on victim networks. [5] It urged network defenders to follow Microsoft’s remediation guidance and a joint FBI and CISA advisory. [5]
Notice followed the 2016 rule’s pattern. The FBI said it was attempting to notify all owners or operators of computers from which it removed web shells. [5] For victims with public contact information, it would email from an official FBI.gov account. [5] For others, it would email providers, such as internet service providers, believed to have contact information and ask them to notify the victim. [5]
On August 29, 2023, the department announced an operation against the Qakbot botnet. [6] It said the FBI had identified over 700,000 computers worldwide, including more than 200,000 in the United States, that appeared infected. [6] It said the FBI redirected botnet traffic through FBI servers, which told infected computers to download a law enforcement file that uninstalled the malware. [6]
The department again described limits. It said the action was limited to information the Qakbot actors had installed. [6] It did not remediate other malware or involve access to or modification of owners’ and users’ information. [6] The FBI worked with CISA and private partners on victim notification and remediation. [6]
On January 14, 2025, the department announced a court-authorized operation that deleted PlugX malware from approximately 4,258 U.S.-based computers and networks. [7] It said it had obtained the first of nine warrants in the Eastern District of Pennsylvania in August 2024, and that the last expired on January 3, 2025. [7]
It said the FBI had tested the deletion commands and determined that they did not otherwise affect the computers’ legitimate functions or collect content. [7] It said the FBI was providing notice to U.S. owners through their internet service providers. [7] It encouraged anti-virus software and security updates to prevent reinfection. [7]
The department’s announcements we opened did not say which venue provision each warrant relied on. [5] [6] [7] We do not claim they used paragraph (b)(6). What the 2021 and 2025 announcements show is the pattern the 2016 posts described: action on victims’ machines under court authority and notice afterward through email or providers, with testing beforehand described in 2025. [3] [5] [7]
What held up from 2016 is the department’s account of limits. The 2021 announcement said the operation patched no vulnerabilities and removed no other tools, and the 2023 announcement said the action did not remediate other malware. [5] [6] The 2025 announcement urged security updates to prevent reinfection. [7] What the 2016 posts we read said little about was the notice channel. The 2021 and 2025 operations relied on emails and internet service providers to reach owners. [5] [7]
In our view, that makes the organization’s contact hygiene part of its incident response. A notice that reaches a stale address or an unprepared provider arrives too late to help. The rule required reasonable efforts by the government. [8] The rest depends on whether the organization can be found.
Spotted an error? Tell us and we will check it and correct the article where needed.