T1070.004: File Deletion
Adversaries may delete files left behind by the actions of their intrusion activity. Malware, tools, or other non-native files dropped or created on a system by an adversary (ex: Ingress Tool Transfer) may leave traces to indicate to what was done within a network and how. Removal of these files can occur during an intrusion, or as part of a post-intrusion process to minimize the adversary's footprint.
There are tools available from the host operating system to perform cleanup, but adversaries may use other tools as well.[1] Examples of built-in Command and Scripting Interpreter functions include del on Windows, rm or unlink on Linux and macOS, and `rm` on ESXi.
Analyst context for executives and security teams
File Deletion matters because it is a cleanup behavior: adversaries may remove malware, transferred tools, scripts, or other files that would otherwise explain what happened. For leaders, the risk is not just the deleted file; it is the loss of incident evidence that slows scoping, containment, legal reporting, and recovery decisions across Windows, Linux, macOS, and ESXi environments.
Executive priority
Treat this as an evidence-preservation and resilience issue. Security leaders should ask whether endpoint, server, and ESXi activity is logged centrally enough that local file cleanup does not erase the investigation trail. The relationship context shows this behavior appears across many ATT&CK-documented groups and campaigns, including espionage, supply-chain, critical infrastructure, and network-device-focused activity, so coverage should be prioritized for high-value systems, administrative hosts, and platforms where IR visibility is thin.
Technical view
SOC and IR teams should validate visibility into file deletion paired with process and command execution on ESXi, Linux, macOS, and Windows. The technique explicitly includes built-in commands such as Windows del, Linux/macOS rm or unlink, and ESXi rm, as well as possible non-native cleanup tools such as SDelete. Because MITRE provides no official detection text for this object, detections should be behavior-based: suspicious deletion of recently dropped tools, scripts, payloads, staging directories, or intrusion artifacts, especially when adjacent to Ingress Tool Transfer or Command and Scripting Interpreter activity. The related DET0140 detection strategy, Behavioral Detection of Malicious File Deletion, is the ATT&CK-provided relationship to use as a detection engineering anchor.
Likely telemetry
- Endpoint/file-system deletion events, including path, user, timestamp, and process context
- Process creation telemetry for del, rm, unlink, ESXi rm, SDelete, and shell/script interpreters
- Command-line and script execution logs from Windows, Linux, macOS, and ESXi where available
- EDR or host audit records tying deleted files to parent processes and users
- Centralized logs that survive local file removal
Detection direction
- Confirm that deletion events are correlated with process execution, not reviewed as standalone noise.
- Tune for suspicious context: deletion of recently created or transferred files, unusual cleanup from user-writable or temporary locations, and deletion shortly after command or script execution.
- Baseline legitimate administrative cleanup, package management, log rotation, and deployment activity to reduce false positives.
- Pay special attention to ESXi and non-Windows systems, where endpoint telemetry and command auditing are often less mature.
- Use relationship context from T1070 Indicator Removal to look for selective artifact removal rather than only bulk wiping.
Mitigation priorities
- Prioritize centralized collection and retention of host, process, and file activity so local deletion does not remove investigative evidence.
- Restrict and monitor use of non-native cleanup utilities and administrative deletion tooling where business operations allow.
- Apply least-privilege controls so ordinary accounts cannot remove high-value security artifacts or administrative tooling traces outside their role.
- Build IR playbooks that preserve volatile evidence and correlate file creation, transfer, execution, and deletion timelines.
- Validate coverage on Windows, Linux, macOS, and ESXi separately rather than assuming one platform’s controls generalize to the others.
Analyst notes and limits
This is a stealth sub-technique of T1070 Indicator Removal and supersedes the revoked T1107 File Deletion technique. The official ATT&CK object is clear about the behavior and platforms but does not provide a detection analytic. Relationship context indicates broad ATT&CK-documented usage across campaigns and groups, but that should be used for prioritization and threat modeling, not as proof of current activity in any specific environment.
No official MITRE detection or mitigation text was supplied for this object. Telemetry and control recommendations are defensive inferences from the official description, platforms, command examples, external reference, and ATT&CK relationships. Local logging architecture, retention, endpoint coverage, and administrative baselines are required to determine actual detection coverage.
File Deletion
Adversaries may delete files left behind by the actions of their intrusion activity. Malware, tools, or other non-native files dropped or created on a system by an adversary (ex: Ingress Tool Transfer) may leave traces to indicate to what was done within a network and how. Removal of these files can occur during an intrusion, or as part of a post-intrusion process to minimize the adversary's footprint.
There are tools available from the host operating system to perform cleanup, but adversaries may use other tools as well.[1] Examples of built-in Command and Scripting Interpreter functions include del on Windows, rm or unlink on Linux and macOS, and `rm` on ESXi.
How security teams should use this page
Treat this object as behavior context, not an attribution claim. Validate the related groups, software, data sources, and mitigations against official ATT&CK relationships and your own telemetry before making control-coverage decisions.
Related techniques
This mirrors the MITRE pattern of making group, software, campaign, and technique relationships scannable. Relationship notes come from mirrored ATT&CK relationship text when available.
| Domain | ID | Name | Relationship / procedure |
|---|---|---|---|
| Enterprise | T1070 | Indicator Removal | This object subtechnique of Indicator Removal. |
Groups, software, and campaigns
G0143: Aquatic Panda
Aquatic Panda is a suspected China-based threat group with a dual mission of intelligence collection and industrial espionage. Active since at least May 2020, Aquatic Panda has primarily targeted entities in the telecommunications, technology, and government sectors.[1]
G0051: FIN10
G0045: menuPass
menuPass is a threat group that has been active since at least 2006. Individual members of menuPass are known to have acted in association with the Chinese Ministry of State Security's (MSS) Tianjin State Security Bureau and worked for the Huaying Haitai Science and Technology Development Company.[1][2]
menuPass has targeted healthcare, defense, aerospace, finance, maritime, biotechnology, energy, and government sectors globally, with an emphasis on Japanese organizations. In 2016 and 2017, the group is known to have targeted managed IT service providers (MSPs), manufacturing and mining companies, and a university.[3][4][5][6][7][1][2]
G0060: BRONZE BUTLER
BRONZE BUTLER is a cyber espionage group with likely Chinese origins that has been active since at least 2008. The group primarily targets Japanese organizations, particularly those in government, biotechnology, electronics manufacturing, and industrial chemistry.[1][2][3]
G0081: Tropic Trooper
Tropic Trooper is an unaffiliated threat group that has led targeted campaigns against targets in Taiwan, the Philippines, and Hong Kong. Tropic Trooper focuses on targeting government, healthcare, transportation, and high-tech industries and has been active since 2011.[1][2][3]
G0139: TeamTNT
TeamTNT is a threat group that has primarily targeted cloud and containerized environments. The group as been active since at least October 2019 and has mainly focused its efforts on leveraging cloud and container resources to deploy cryptocurrency miners in victim environments.[1][2][3][4][5][6][7][8][9]
G0082: APT38
APT38 is a North Korean state-sponsored threat group that specializes in financial cyber operations; it has been attributed to the Reconnaissance General Bureau.[1] Active since at least 2014, APT38 has targeted banks, financial institutions, casinos, cryptocurrency exchanges, SWIFT system endpoints, and ATMs in at least 38 countries worldwide. Significant operations include the 2016 Bank of Bangladesh heist, during which APT38 stole $81 million, as well as attacks against Bancomext [2] and Banco de Chile [2]; some of their attacks have been destructive.[1][2][3][4]
North Korean group definitions are known to have significant overlap, and some security researchers report all North Korean state-sponsored cyber activity under the name Lazarus Group instead of tracking clusters or subgroups.
G0053: FIN5
FIN5 is a financially motivated threat group that has targeted personally identifiable information and payment card information. The group has been active since at least 2008 and has targeted the restaurant, gaming, and hotel industries. The group is made up of actors who likely speak Russian. [1] [2] [3]
G1040: Play
Play is a ransomware group that has been active since at least 2022 deploying Playcrypt ransomware against the business, government, critical infrastructure, healthcare, and media sectors in North America, South America, and Europe. Play actors employ a double-extortion model, encrypting systems after exfiltrating data, and are presumed by security researchers to operate as a closed group.[1][2]
G1043: BlackByte
BlackByte is a ransomware threat actor operating since at least 2021. BlackByte is associated with several versions of ransomware also labeled BlackByte Ransomware. BlackByte ransomware operations initially used a common encryption key allowing for the development of a universal decryptor, but subsequent versions such as BlackByte 2.0 Ransomware use more robust encryption mechanisms. BlackByte is notable for operations targeting critical infrastructure entities among other targets across North America.[1][2][3][4][5]
G1003: Ember Bear
Ember Bear is a Russian state-sponsored cyber espionage group that has been active since at least 2020, linked to Russia's General Staff Main Intelligence Directorate (GRU) 161st Specialist Training Center (Unit 29155).[1] Ember Bear has primarily focused operations against Ukrainian government and telecommunication entities, but has also operated against critical infrastructure entities in Europe and the Americas.[2] Ember Bear conducted the WhisperGate destructive wiper attacks against Ukraine in early 2022.[3][4][1] There is some confusion as to whether Ember Bear overlaps with another Russian-linked entity referred to as Saint Bear. At present available evidence strongly suggests these are distinct activities with different behavioral profiles.[2][5]
G1054: MirrorFace
MirrorFace is a People's Republic of China (PRC)-aligned cyberespionage actor believed to be a subgroup under the menuPass umbrella based on targeting, tools, and infrastructure overlaps. MirrorFace has been active since at least 2019, at first exclusively targeting Japanese organizations across the media, defense, diplomatic, financial, manufacturing, and academic sectors. Subsequent MirrorFace operations included targets in Central Europe and featured use of LODEINFO, HiddenFace, and UPPERCUT malware.[1][2][3][4][5][6]
S0164: TDTESS
TDTESS is a 64-bit .NET binary backdoor used by CopyKittens. [1]
S9020: LODEINFO
LODEINFO is a fileless backdoor malware first identified in 2020 that has been used by actors including MirrorFace, primarily against media, diplomatic, governmental, and public sector organizations in Japan.[1][2][3]
S0395: LightNeuron
LightNeuron is a sophisticated backdoor that has targeted Microsoft Exchange servers since at least 2014. LightNeuron has been used by Turla to target diplomatic and foreign affairs-related organizations. The presence of certain strings in the malware suggests a Linux variant of LightNeuron exists.[1]
S1150: ROADSWEEP
ROADSWEEP is a ransomware that was deployed against Albanian government networks during HomeLand Justice along with the CHIMNEYSWEEP backdoor.[1]
S0654: ProLock
S1212: RansomHub
RansomHub is a ransomware-as-a-service (RaaS) offering with Windows, ESXi, Linux, and FreeBSD versions that has been in use since at least 2024 to target organizations in multiple sectors globally. RansomHub operators may have purchased and rebranded resources from Knight (formerly Cyclops) Ransomware which shares infrastructure, feature, and code overlaps with RansomHub.[1][2]
S0354: Denis
S0448: Rising Sun
Rising Sun is a modular backdoor that was used extensively in Operation Sharpshooter between 2017 and 2019. Rising Sun infected at least 87 organizations around the world, including nuclear, defense, energy, and financial service companies. Security researchers assessed Rising Sun included some source code from Lazarus Group's Trojan Duuzer.[1]
S0593: ECCENTRICBANDWAGON
ECCENTRICBANDWAGON is a remote access Trojan (RAT) used by North Korean cyber actors that was first identified in August 2020. It is a reconnaissance tool--with keylogging and screen capture functionality--used for information gathering on compromised systems.[1]
S0370: SamSam
S0390: SQLRat
S1027: Heyoka Backdoor
Heyoka Backdoor is a custom backdoor--based on the Heyoka open source exfiltration tool--that has been used by Aoqin Dragon since at least 2013.[1][2]
C0022: Operation Dream Job
Operation Dream Job was a cyber espionage operation likely conducted by Lazarus Group that targeted the defense, aerospace, government, and other sectors in the United States, Israel, Australia, Russia, and India. In at least one case, the cyber actors tried to monetize their network access to conduct a business email compromise (BEC) operation. In 2020, security researchers noted overlapping TTPs, to include fake job lures and code similarities, between Operation Dream Job, Operation North Star, and Operation Interception; by 2022 security researchers described Operation Dream Job as an umbrella term covering both Operation Interception and Operation North Star.[1][2][3][4]
C0029: Cutting Edge
Cutting Edge was a campaign conducted by suspected China-nexus espionage actors, variously identified as UNC5221/UTA0178 and UNC5325, that began as early as December 2023 with the exploitation of zero-day vulnerabilities in Ivanti Connect Secure (previously Pulse Secure) VPN appliances. Cutting Edge targeted the U.S. defense industrial base and multiple sectors globally including telecommunications, financial, aerospace, and technology. Cutting Edge featured the use of defense evasion and living-off-the-land (LoTL) techniques along with the deployment of web shells and other custom malware.[1][2][3][4][5]
C0006: Operation Honeybee
Operation Honeybee was a campaign that targeted humanitarian aid and inter-Korean affairs organizations from at least late 2017 through early 2018. Operation Honeybee initially targeted South Korea, but expanded to include Vietnam, Singapore, Japan, Indonesia, Argentina, and Canada. Security researchers assessed the threat actors were likely Korean speakers based on metadata used in both lure documents and executables, and named the campaign "Honeybee" after the author name discovered in malicious Word documents.[1]
All related ATT&CK context
Object version and sync metadata
The fields below describe the current mirrored snapshot. When Glexia retains multiple ATT&CK source imports, you can open the table to compare the same object across releases (hashes and MITRE timestamps). For MITRE’s own release notes and roadmap, see ATT&CK resources — Updates .
Imported snapshots across ATT&CK releases (1)
| Release | Bundle imported | Object version | Modified | Status | Raw hash |
|---|---|---|---|---|---|
| 19.1 | 2.0 | Current bundle | e2b06aca3e3f… |
Mirrored ATT&CK source object
The raw object is retained through the mirrored ATT&CK source bundle and object hash. The raw endpoint returns the exact object from the mirrored bundle when available.
External references and citations
MITRE external references are preserved separately from Glexia analysis so citations remain traceable to their original source records.
-
[1]
Microsoft SDelete July 2016
Russinovich, M. (2016, July 4). SDelete v2.0. Retrieved February 8, 2018.
Open source URL -
[2]
mitre-attack T1070.004Open source URL
Source: MITRE ATT&CK®. © 2026 The MITRE Corporation. This work is reproduced and distributed with the permission of The MITRE Corporation. MITRE ATT&CK and ATT&CK are registered trademarks of The MITRE Corporation. Glexia is not affiliated with or endorsed by MITRE.