A.8.7 Technological
Protection against malware
Structured attributes from ISO/IEC 27002:2022 — control type · CIA properties · cybersecurity concept · operational capability · security domain. What do these mean?
Mapped NIST 800-53 r5 controls (17)
Our reading (mostly llm_verified) of how this ISO control and each NIST 800-53 control relate. Not an ISO or NIST product.
Direction: ← other covers this;
→ this covers other (F/M/P = full / mostly /
partial). gov = governs / implements (a mandate, not coverage).
Why these map — rationale
- SI-3mostlycovers — A.8.7's requirement to protect assets against malware is largely accounted for by SI-3's specific mechanisms, detection/eradication, updates, and scanning at entry/exit points, but a residual of A.8.7 (e.g., user awareness, non-technical controls, or protection of assets outside system boundaries) remains uncovered.
- CM-7partialcovers — A.8.7's malware-specific protections (e.g. detection, prevention, updates) address only the malware-related slice of CM-7's broader least-functionality and prohibited-software requirements; the bulk of CM-7 (non-malware ports, protocols, services, and general configuration) sits outside A.8.7
Aligned NIST CSF 2.0 outcomes (18)
NIST CSF 2.0 outcomes this ISO control aligns with — our analysis (mostly llm_verified).
Direction: ← other covers this;
→ this covers other (F/M/P = full / mostly /
partial). gov = governs / implements (a mandate, not coverage).
Why these map — rationale
- PR.PS-05fullcovers — The ISO control's emphasis on application allowlisting, blocklisting, and preventing unauthorized software execution directly fulfills the CSF outcome of stopping installation and execution of unauthorized software.
Related OWASP ASVS 5.0 requirements (12)
Application-security verification requirements (OWASP ASVS 5.0) this ISO control aligns with; links open the ASVS chapter. Our analysis (mostly llm_verified) — many ISO controls have no ASVS counterpart.
Direction: ← other covers this;
→ this covers other (F/M/P = full / mostly /
partial). gov = governs / implements (a mandate, not coverage).
Related weaknesses / CWE (2404)
Weakness classes this ISO control helps prevent or mitigate — our analysis (mostly llm_verified).
Direction: ← other covers this;
→ this covers other (F/M/P = full / mostly /
partial). gov = governs / implements (a mandate, not coverage).
Why these map — rationale
- CWE-507mostlyfinds — Malware protection directly detects and blocks Trojan Horse code before execution.
- CWE-507mostlyremoves — A.8.7's malware detection and repair software (item f), when applied to a system that previously lacked it, scans for and removes existing malicious code including Trojans; the remainder is bounded and nameable — Trojans whose hidden payload is well-integrated into legitimate functionality and evades signature, heuristic and behavioral detection, which is precisely the 'hidden from normal operation' property CWE-507 names.
- CWE-508mostlyblocks — A.8.7 is the dedicated malware control and its mechanisms — detection/repair software, application allowlisting, network/email/web scanning, and defense-in-depth placement — directly block execution of non-replicating malicious code; the bounded remainder is sophisticated or zero-day malware that evades signature detection and exploits allowed applications (living-off-the-land), where allowlisting alone does not stop abuse of legitimate tools.
- CWE-508mostlyremoves — A.8.7's malware detection and repair software (item f) is specifically designed to scan systems, find existing malicious code, and remove or quarantine it — applied where it was not applied before, it removes most instances of CWE-508 (non-replicating malicious code already present on a system). The bounded remainder is sophisticated, novel, or evasive malware that evades signature-based detection, which the clause itself addresses via items g.2 (evasive techniques), n (new malware intelligence), and o (qualified sources), but cannot fully close.
- CWE-508mostlyfinds — A.8.7's detection and repair software, regular automated validation of system content (8.7d), and scanning of files/email/web (8.7f) are explicitly designed to discover malicious code residing on systems, which is exactly what CWE-508 (non-replicating malicious code such as trojans and backdoors) describes; the bounded remainder is sophisticated or novel malware that evades signature and behavioural detection.
- CWE-509mostlyblocks — A.8.7's layered mechanisms — detection-and-repair scanning at gateways and endpoints, application allowlisting, blocklisting of malicious sites, and isolation of high-risk environments — are precisely aimed at stopping replicating malicious code from executing and propagating, which is blocking exploitation of the weakness; the bounded remainder is evasive and zero-day malware that evades signature-based detection and reaches execution before the control can act, which the clause itself acknowledges ('detection and repair software alone is not usually adequate').
- CWE-509mostlyfinds — A.8.7's malware detection and repair software, email/web scanning, and blocklisting are specifically designed to find viruses and worms — the primary purpose of the control is discovering instances of replicating malicious code in the operational environment, and signature plus heuristic coverage reaches the bulk of known CWE-509 instances; the bounded remainder is zero-day and sophisticated evasion malware (polymorphic code, encrypted payloads, targeted threats) that signature-based detection misses.
- CWE-509mostlyremoves — A.8.7(f) explicitly mandates 'malware detection and repair software' — the repair function is precisely the removal of malicious code already present on a system, which is the CWE-509 instance. Well-implemented anti-malware with current signatures detects and removes most known viruses and worms; the bounded remainder is sophisticated, evasive, or zero-day malware that evades signature-based detection.
- CWE-512mostlyfinds — A.8.7's malware detection and repair software (guidance item f) is specifically designed to detect malicious software including spyware, and modern anti-malware products treat spyware as a recognized category; a well-implemented programme with signature, heuristic and behavioural detection finds the bulk of spyware instances, with a bounded remainder of novel or sophisticated variants that evade detection until signatures are developed.
- CWE-512mostlyblocks — A.8.7's malware detection, application allowlisting, and malicious-website blocklisting directly target spyware as a malware class, blocking most instances from executing or persisting; the bounded remainder is zero-day or evasion-technique spyware and spyware bundled with allowlisted software, which detection signatures and allowlists miss.
- CWE-1338partialblocks — A.8.7 blocks the malware-based exploitation path for CWE-1338 (the CWE explicitly identifies malicious software as the attack vector), but the clause itself states detection software alone is not usually adequate, so some malware reaches execution; the underlying hardware weakness (missing thermal sensors, throttling, cooling) is entirely unaddressed, and non-malicious software running in high-power modes could still trigger it.
- CWE-356partialblocks — A.8.7's malware detection and repair software, scanning files and email attachments before use (items f.1, f.2), blocks exploitation of CWE-356 in the subset of cases where the unsafe action involves malware (e.g., executing a malicious attachment, opening a macro-laden document), but CWE-356 also covers non-malware unsafe actions such as certificate-acceptance warnings and setuid extraction warnings, which A.8.7 does not reach.
- CWE-426partialblocks — A.8.7's application allowlisting (guidance item a) can prevent the malicious binary an attacker plants in the search path from being installed or executed, which blocks a meaningful slice of CWE-426 exploitation; however, the weakness is a code-level defect in how the target program resolves paths, and allowlisting does not reach cases where the attacker abuses existing legitimate binaries in writable directories, hijacks DLLs, or controls the search path itself through other means.
- CWE-427partialblocks — A.8.7's application allowlisting (8.7a) and malware detection (8.7f) can block exploitation of an uncontrolled search path by preventing the attacker's planted DLL/executable from running or detecting it before load, but the code-level weakness itself is untouched and allowlisting does not cover configuration-file or data-file search path elements, leaving a nameable remainder.
- CWE-428partialblocks — A.8.7 can block some exploitation by allowlisting or quarantining attacker-supplied executables, but it does not neutralize the unquoted-path weakness and can miss or permit non-malicious payloads.
- CWE-434partialblocks — A.8.7 can block exploitation of CWE-434 when uploaded dangerous files contain malware that its scanning detects before use, but because it does not restrict file types itself and malware detection misses unknown or custom scripts, only a minority of instances are stopped.
- CWE-494partialblocks — A.8.7's malware detection and repair software (guidance f) scans data received over networks and downloads for known malware before use, which blocks some exploitation of CWE-494 by catching malicious payloads in transit. However, the underlying weakness — the product's own lack of integrity verification on downloaded code — remains present and exploitable by zero-day or signature-evading malware, and by adversary-in-the-middle substitution of legitimate-looking updates, which signature-based scanning does not reach.
- CWE-507partialblocks — A.8.7's allowlisting, malware detection, and email/web scanning block many Trojans delivered as standalone malicious files, but CWE-507 specifically describes hidden malicious code embedded within software that appears legitimate, and that class is largely outside the clause's reach — allowlisting passes allowed software through, and detection can be evaded by code hidden inside a trusted product.
- CWE-511partialremoves — A.8.7's malware detection and repair software can remove some logic/time bombs that arrive as separate malicious files with detectable signatures, but the majority of instances are embedded in legitimate application code (first-party by a malicious developer or third-party via supply chain) where AV cannot remove the defect without breaking the host application, leaving a nameable remainder.
- CWE-511partialfinds — A.8.7's anti-malware programme (signatures, heuristics, behavioural monitoring, sandboxing, plus the item (d) automated validation of software/data content) finds some logic/time bombs — known payloads, obvious trigger patterns, and unauthorized amendments — but the bulk of CWE-511 instances are custom dormant code embedded in otherwise legitimate software that evades signature and heuristic detection, leaving a nameable remainder.
- CWE-512partialremoves — A.8.7's malware detection and repair software (item f) removes spyware instances from systems where they are flagged as malware, but CWE-512 is a design weakness in a product's information-collection behavior, not necessarily malware that infected a legitimate system; many products exhibiting CWE-512 (e.g., legitimate software with covert telemetry) would not be detected or removed, and the underlying design flaw in the product itself is untouched by removal from any single system.
- CWE-553partialblocks — A.8.7's allowlisting and pre-use malware scanning can block execution of uploaded web shells, but it does not reliably prevent exploitation of authorized, benign, or evasive command-shell files exposed in a web-accessible directory.
- CWE-69partialfinds — Malware-protection mechanisms may scan or block known ADS-based threats, yet coverage is limited to signature or behavior heuristics rather than all ADS handling flaws.
- CWE-829partialblocks — A.8.7's malware detection (8.7f) and web scanning (8.7f.3) can block exploitation when the included untrusted functionality carries malware, and allowlisting (8.7a) can prevent some unauthorized included code from running, but the weakness's core exploitation paths — DOM-based XSS, excessive privileges granted by the included code, vulnerabilities within the included functionality itself, and compromise of the untrusted source between scans — are not addressed by malware-focused controls.
- CWE-98partialblocks — A.8.7 can block known malicious remote sources and detected malware content from being included, but it does not address the underlying filename-control flaw and leaves local file inclusion or unknown remote payloads exploitable.
Mitigated MITRE ATT&CK techniques (3898)
Adversary techniques (MITRE ATT&CK Enterprise) this ISO control helps mitigate; links open attack.mitre.org. Our analysis (mostly llm_verified).
Direction: ← other covers this;
→ this covers other (F/M/P = full / mostly /
partial). gov = governs / implements (a mandate, not coverage).
Why these map — rationale
- T1036.007mostlyprevents — A.8.7 prevents T1036.007 through multiple mechanisms: email attachment scanning (f.2) catches the malicious executable regardless of its displayed extension, application allowlisting (a) blocks unauthorized executables from running, and malware detection identifies the file by its actual content/type. The bounded remainder is novel malware evading signature detection and allowlisting exceptions, plus the user-deception aspect itself which the control does not address.
- T1055mostlydetects — A.8.7 explicitly requires malware detection software, regular automated scans of systems/data/email/web content, anomaly detection via validation of software/content, and monitoring for unauthorized changes or files, all of which surface process injection as anomalous/malicious behavior in flight on monitored platforms.
- T1055.001mostlydetects — A.8.7 explicitly requires malware detection/repair software, regular automated scans of systems/data/email/webpages, collection of new malware intel, and placement based on risk (including defence-in-depth at endpoints/servers/gateways), which surfaces in-process DLL injection as anomalous/malicious code or behaviour on Windows platforms.
- T1055.002mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans (including of received data, email attachments, downloads, and webpages), plus awareness training, vulnerability management, and monitoring for unapproved files/amendments, all of which surface PE injection as anomalous or malicious behavior in flight on monitored Windows systems.
- T1055.003mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans of systems, received data, email/IM attachments, webpages, and storage media (including at multiple points like gateways, endpoints, and servers per defence-in-depth), which surfaces the anomalous code/memory writes and process anomalies characteristic of thread execution hijacking as malware.
- T1055.004mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular scans of systems, received data, email/IM attachments, downloads, and webpages, plus awareness and procedures to identify/report malware; this surfaces APC injection (including Early Bird and AtomBombing variants) when it manifests as anomalous/malicious code or files in processes, with the bounded remainder being highly evasive in-memory-only cases that avoid observable artifacts or scans.
- T1056.001mostlydetects — A.8.7's malware detection and repair software (f, g) and application allowlisting (a) are designed to surface malicious software including keyloggers, and modern EDR/AV products specifically detect keyloggers via behavioral signatures on API hooking, custom drivers, and registry modifications; the bounded remainder is hardware-based keyloggers reading raw keystroke data from the hardware buffer and sophisticated kernel-level implementations that the control itself acknowledges via its evasive-techniques clause (g.2).
- T1056.004mostlydetects — A.8.7 explicitly requires malware detection/repair software, regular automated validation/scans of systems/data/email/web content, collection of new malware intel, and defence-in-depth placement (gateways, endpoints, servers) that surfaces hooking-based credential stealers as malware or anomalous files/behaviour; the named remainder is non-malware implementations or evasive encrypted/zero-day hooks that slip past signature/heuristic detection.
- T1059.005mostlydetects — A.8.7 mandates malware detection and repair software with scanning of email attachments, downloads, and webpages, which directly surfaces VB/VBA/VBScript payloads in flight; the bounded remainder is the evasive techniques the clause itself names (encrypted files, obfuscation) that defeat signature-based detection.
- T1059.007mostlydetects — A.8.7 explicitly requires malware detection/repair software, regular automated scans of received data/email/webpages, collection of new malware intel, and placement based on risk (including defense-in-depth at gateways/endpoints/servers), which surfaces JS-based execution payloads or behaviors in flight or at delivery for the bulk of common cases across platforms.
- T1059.010mostlydetects — A.8.7 mandates malware detection software scanning files, email attachments, downloads and webpages, plus application allowlisting that surfaces unauthorized script execution; compiled AHK/AutoIT payloads and known-malicious scripts are caught, with the bounded remainder being novel or obfuscated scripts that evade signature-based detection because the underlying tools are legitimate.
- T1091mostlydetects — A.8.7(f) and (f.1) explicitly require scanning 'electronic storage media' for malware, which covers the bulk of T1091's file-based payloads on USB drives and similar media when inserted into endpoints; the bounded remainder is firmware-level implants and evasion techniques that bypass signature-based file scanning, plus air-gapped systems where signature freshness lags.
- T1091mostlyprevents — A.8.7 directly addresses removable media through (e) protective measures for external media, (f.1) scanning data received via electronic storage media, and (a) application allowlisting that blocks Autorun and unauthorized executable execution; however, firmware-level manipulation of the media itself and social engineering that tricks users into executing disguised files on air-gapped systems remain outside the control's reach.
- T1127.001mostlydetects — A.8.7 explicitly requires regular automated validation of software/content, scanning of received data/email/webpages, collection of new malware intel from reputable sources, and risk-based placement of detection tools (including at endpoints, servers, gateways and email), which surfaces MSBuild abuse when it triggers observable indicators such as anomalous inline task compilation or known malicious patterns.
- T1127.002mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans of received data, email/IM attachments, downloads, webpages, external files, and media (points d, f, g), which surfaces ClickOnce-based proxy execution when the .application/.appref-ms or associated malicious payload is delivered or executed on monitored Windows systems.
- T1129mostlydetects — A.8.7 mandates malware detection and repair software that scans files, email, webpages, and storage media, and shared modules loaded from arbitrary paths are files on disk or in transit that signature-based and heuristic scanning can surface before or at load time.
- T1137.001mostlydetects — A.8.7 mandates malware detection and repair software with regular scanning of files, email attachments, and storage media, which would surface a malicious Office template macro on disk or during a scan of the template file location.
- T1189mostlydetects — A.8.7 explicitly enumerates detection mechanisms aimed at this technique — scanning webpages for malware when accessed (f.3), detecting use of known or suspected malicious websites via blocklisting (b), and defence-in-depth scanning at network gateways (g.1) — so the control surfaces drive-by compromise attempts at multiple points; the bounded remainder is evasion through legitimate-but-compromised sites, zero-day browser exploits, and encrypted delivery channels that the clause itself names (g.2) as a concern.
- T1189mostlyprevents — A.8.7 directly addresses drive-by compromise through multiple specific mechanisms: blocklisting malicious websites (b), reducing exploitable vulnerabilities via patching (c), scanning webpages for malware when accessed (f.3), and application allowlisting to constrain post-exploitation payloads (a). The remainder is nameable: legitimate-but-compromised sites evade blocklists, zero-day browser exploits bypass signature-based scanning, and browsers must remain functional for normal operations — so the technique cannot be fully prevented even with the control's defence-in-depth posture.
- T1203mostlydetects — A.8.7 mandates malware detection and repair software that scans files, email, webpages, and storage media before use, and T1203's delivery mechanisms—malicious documents, browser exploits, and third-party application files—all pass through these scanning points, leaving only encrypted or zero-day exploits as a bounded remainder.
- T1204.002mostlydetects — A.8.7's scanning requirements (f.1, f.2, f.3) and detection-and-repair software are designed precisely to surface malicious files before or at execution, and the clause explicitly addresses evasive techniques (g.2); the bounded remainder is encrypted/obfuscated payloads and zero-days that evade signature and heuristic detection.
- T1204.002mostlyprevents — A.8.7 stacks multiple prevention layers against T1204.002: pre-delivery scanning of email attachments and downloads (f.2), webpage scanning (f.3), application allowlisting that blocks execution even when a user opens the file (a), and user awareness/training (m) that lowers the opening rate. Together these catch the technique at several points, but the technique is fundamentally user-driven and adversaries use obfuscation, masquerading and novel payloads that evade signature detection, while allowlisting can be bypassed via LOLbins and signed binaries — so prevention is strong but not complete.
- T1204.002mostlyresponds — A.8.7 explicitly requires business continuity plans, isolating high-risk environments, procedures for reporting/recovering from malware attacks, and protective measures during maintenance that bypass normal controls, all of which act on a realized malicious-file execution to contain, eradicate, and restore after the technique has run.
- T1204.005mostlydetects — A.8.7 mandates malware detection and repair software that scans files, email attachments, downloads, and webpages before use, and malicious libraries obtained from package managers or repositories are files obtained from external networks that would be scanned at multiple points (gateway, endpoint) as they enter the environment.
- T1218.001mostlydetects — A.8.7 explicitly mandates scanning email attachments, downloads, and webpages for malware at multiple points (gateway, endpoint, server), and CHM files delivered as email attachments or downloads fall squarely within that scope; defense-in-depth placement and consideration of evasive techniques (g) further strengthen coverage, but the clause itself acknowledges detection software alone is not usually adequate and novel or obfuscated CHM payloads can evade signature-based scanning.
- T1218.005mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular scans of networks, email/IM attachments, downloads, webpages and storage media (f, g), plus awareness and procedures to report malware (l,m); these surface mshta.exe abuse of .hta/.sct payloads (a common malware vector) when it reaches monitored entry points or exhibits anomalous behavior, though placement/config is risk-based and evasive/encrypted cases leave a bounded remainder.
- T1218.007mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans of files, email attachments, downloads, webpages and storage media (including at multiple points such as servers, endpoints and network entry), which surfaces msiexec abuse when it delivers or executes known/suspicious malware payloads; the remainder is evasive/unknown variants that bypass signature-based detection.
- T1218.009mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans of received data, email/IM attachments, downloads, webpages and storage media (including at multiple points such as servers, endpoints and network entry), plus awareness/training on identifying malware-infected files; this surfaces the signed-trusted-utility proxy execution of malicious code that is the technique's core mechanism, with the bounded remainder being evasive/encrypted delivery that the clause itself notes may require additional risk-based placement.
- T1218.010mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular scans of data, email/IM attachments, downloads, webpages and storage media (including at multiple points such as gateways, servers and endpoints), plus awareness of new malware and verification of threat intel, which surfaces Regsvr32 abuse when it loads or executes malicious code or scriptlets.
- T1505.003mostlydetects — A.8.7 mandates malware detection and repair software that scans files, email, webpages, and storage media, and web shells are files placed on a web server that such scanning can surface on access or during regular automated validation, with the bounded remainder being web shells delivered through encrypted channels or evasive techniques that the control itself acknowledges may bypass detection.
- T1546.010mostlydetects — A.8.7 explicitly requires regular automated validation of system software/content, malware detection software with scans of files/downloads/webpages at multiple layers (gateway/endpoint/server), collection of new malware intel, and awareness-driven reporting, all of which surface AppInit DLL abuse as anomalous or malicious code on Windows.
- T1547.001mostlydetects — A.8.7 mandates malware detection and repair software (f), rules detecting unauthorized software (a), and automated validation investigating unauthorized amendments (d); modern endpoint protection products commonly flag registry run key and startup folder modifications as persistence indicators, so the technique is broadly detectable under the clause, but the clause's primary focus is malware payloads rather than persistence mechanisms, and evasion via masquerading or referencing legitimate programs leaves a bounded remainder.
- T1547.004mostlydetects — A.8.7 mandates regular automated scanning of systems and files, which would surface unauthorized modifications to Winlogon registry keys and the associated DLLs on disk, though it may miss in-memory-only variants or encrypted payloads before execution.
- T1547.010mostlydetects — A.8.7 explicitly requires regular automated validation of software/content, malware detection software with scans (including on boot-time loading paths and system files), anomaly detection for unauthorized files/amendments, and monitoring for new malware indicators, all of which surface the port monitor DLL persistence technique on Windows.
- T1554mostlydetects — A.8.7 mandates regular automated validation of software and data content, scanning for unapproved files or unauthorized amendments, which directly surfaces modifications to host software binaries.
- T1561.001mostlydetects — A.8.7 mandates malware detection and repair software with regular scanning of files, email, webpages, and storage media, which would surface the malware performing the disk wipe in flight, though detection may be bypassed by evasive techniques like encryption or direct disk access drivers.
- T1566mostlyresponds — A.8.7 explicitly requires procedures, responsibilities, training, awareness, business continuity plans and recovery measures to deal with and recover from malware attacks delivered via phishing emails/attachments/links, which matches the `responds` verb once the technique is underway.
- T1566.001mostlydetects — A.8.7(f)(2) explicitly mandates scanning email attachments for malware at multiple layers (email servers, gateways, desktops), which directly detects spearphishing attachments, but the clause itself names evasion techniques (encrypted/password-protected files, encryption protocols) that bypass this detection, preventing a full grade.
- T1566.001mostlyprevents — A.8.7 directly addresses spearphishing attachments through email attachment scanning at gateways and endpoints (f.2, g.1), application allowlisting that prevents payload execution (a), and defense-in-depth placement — but signature-based scanning cannot catch novel or encrypted attachments, the exception process (i) can disable controls, and user execution can still bypass protections, leaving a named remainder.
- T1566.001mostlyresponds — A.8.7 explicitly requires procedures to deal with malware attacks (l), business continuity plans for recovering from them (j), isolating high-risk environments (k), awareness/training on identifying and mitigating infected attachments/emails (m), and scanning email attachments for malware before use at multiple layers (f.2), which together contain, eradicate and recover from a realized spearphishing-attachment event once underway.
- T1569mostlydetects — A.8.7 explicitly requires malware detection/repair software, regular automated scans of systems/data/email/webpages, collection of new malware intel, and awareness of malware indicators; these surface abuse of system services/daemons when it involves known malicious executables, boot persistence, or anomalous service creation — the dominant shapes of T1569 — while non-malware service abuse (purely administrative) sits outside the malware-specific mechanism.
- T1574mostlydetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans of files, email, downloads, webpages and storage media (f, d, g), plus awareness and procedures to report malware (l, m), which surfaces hijacked execution when it manifests as recognizable malware payloads or anomalous files/amendments on the monitored platforms.
- T1574.001mostlydetects — A.8.7 mandates malware detection and repair software with regular scanning of files, email, and web content, which surfaces malicious DLLs on disk or in transit, but the technique's in-memory execution and search-order manipulation can evade signature-based scanning, leaving a bounded remainder.
- T1003partialdetects — A.8.7's malware detection and repair software (clause f) and defense-in-depth placement (clause g) detect known credential-dumping malware such as Mimikatz and its variants, but the control's scope is malware — credential dumping performed via living-off-the-land binaries (procdump, comsvcs.dll, built-in debuggers) or direct memory reads without a malicious binary is outside what malware-focused detection sees, so the technique is caught as a slice rather than as a bounded remainder of a working mechanism.
- T1003partialprevents — A.8.7's application allowlisting (a) and vulnerability reduction (c) can block some credential-dumping tools and exploits, but T1003 frequently uses dual-use admin tools, direct memory access (e.g., LSASS reads), and living-off-the-land binaries that the control's malware-focused mechanisms do not reach.
- T1003partialresponds — A.8.7's malware repair software can quarantine or kill malicious executables performing the dump, but this provides only a partial response because adversaries frequently use legitimate system tools or living-off-the-land binaries (such as ProcDump or Task Manager) which are not eradicated as malware.
- T1003.001partialdetects — A.8.7's malware detection software can identify known credential-dumping tools like mimikatz, but T1003.001 can be executed with legitimate utilities (procdump, comsvcs.dll) and built-in Windows mechanisms (SSP DLLs, silent process exit via WerFault) that malware detection does not flag, so detection covers a slice of instances rather than the technique broadly.
- T1003.002partialdetects — A.8.7's malware detection software detects known credential-dumping tools (Mimikatz, gsecdump, pwdumpx, secretsdump) when they touch disk or exhibit flagged behavior, but the technique can also be performed with built-in Windows utilities (reg.exe) acting legitimately and via in-memory methods that evade signature-based scanning, so only a slice of T1003.002 instances fall within A.8.7's malware-detection scope.
- T1003.002partialprevents — A.8.7's malware detection can catch known credential dumping tools (Mimikatz, gsecdump, secretsdump) and its application allowlisting guidance (a) can block unauthorized post-exploitation tools, but the technique can be executed with built-in Windows utilities (reg.exe) and requires SYSTEM-level access that already implies the adversary has bypassed the control's primary surface, so prevention is a slice rather than a closure.
- T1003.002partialresponds — Malware repair software can quarantine and remove known credential-dumping tools like Mimikatz, but does not contain or eradicate the extraction of the SAM database using native commands like 'reg save'.
- T1003.003partialdetects — A.8.7 mandates malware detection and repair software that scans files and storage media, which would surface NTDS extraction tools if they are known malware, but the technique can also be performed with legitimate built-in Windows tools like ntdsutil.exe that malware scanners may not flag.
- T1003.004partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, and LSA Secrets dumping by Mimikatz is a known malware behaviour that endpoint detection tools can surface, but the control's scanning scope is file- and network-oriented rather than in-memory, so detection depends on whether the specific tool deployed includes memory-scanning capabilities, which the clause does not require.
- T1003.005partialdetects — A.8.7's detection scope is malware — files, email attachments, downloads, webpages, unauthorized software — while T1003.005 is credential access that reads cached credential stores (MS-Cache v2, SSSD/Quest caches) using legitimate or dual-use utilities (Mimikatz, Reg, secretsdump.py, tdbdump), not malware delivery, so the control's scanning and detection mechanisms have no object to act on.
- T1003.006partialdetects — A.8.7 mandates malware detection and repair software that scans files and data, but DCSync abuses a legitimate domain replication API and does not introduce malware, so detection depends on whether the control's monitoring scope extends to anomalous replication traffic rather than on its core malware-scanning function.
- T1003.007partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, but credential harvesting from /proc is an in-memory read by a legitimate process, not a file or network artifact that signature-based or file-scanning malware tools typically observe.
- T1006partialdetects — A.8.7's primary mechanism — malware detection and repair software operating at the file system level — is exactly what T1006 is designed to bypass, but secondary mechanisms in the clause (application allowlisting of tools like vssadmin/esentutl/NinjaCopy, defence-in-depth endpoint behaviour monitoring, and regular automated validation of data content) can surface the technique, leaving a bounded remainder rather than full coverage.
- T1014partialdetects — A.8.7's malware detection and repair software, integrity validation (d), and allowlisting (a) can surface some rootkits, but rootkits are specifically designed to evade these mechanisms by hooking the API calls the detection software relies on, so detection is partial rather than mostly.
- T1014partialprevents — A.8.7's application allowlisting (a) and detection software (f) prevent user-mode rootkits from executing, but kernel-level, bootkit, and firmware rootkits operate below or outside the control's detection surface and bypass these measures by design.
- T1014partialresponds — A.8.7's malware detection and repair software (f) performs removal once malware is found, and (k) calls for isolating environments where catastrophic consequences can occur — both are response actions against malware that has executed. Against rootkits specifically the response is bounded: rootkits hook and hide below the API surface the repair tools rely on, kernel/firmware-resident instances survive standard removal, and the clause's centre of gravity is prevention and detection rather than containment and eradication.
- T1027partialdetects — Regular automated validation of system files and data content helps surface obfuscated or packed malware that would otherwise remain hidden on disk.
- T1027partialprevents — A.8.7 explicitly requires reducing exploitable vulnerabilities, application allowlisting, blocking malicious sites, scanning for malware (including encrypted/obfuscated payloads where detectable), awareness training, and configuration based on risk to stop malware delivery and execution; this stops many T1027 instances that rely on evasion of signature-based or basic controls, but leaves a genuine remainder for advanced evasion (e.g. custom encoding, command obfuscation, or non-malware uses) that the clause does not address.
- T1027partialrecovers — A.8.7 point j) explicitly requires business continuity plans including backups and recovery measures for recovering from malware attacks, which can restore state after an obfuscated payload has executed (e.g. via T1204) and caused impact.
- T1027.001partialdetects — A.8.7's primary mechanism (signature/hash-based malware detection per f) is exactly what binary padding is designed to evade—hash-based blocklists, static AV signatures, and file-size limits are all defeated by the technique, and the control's own guidance (g.2) acknowledges evasive techniques as a known gap. However, the control's other mechanisms—application allowlisting (a), defense in depth (g.1), regular automated validation (d), and behavioral/heuristic components of detection software—can still surface padded malware that signature scanning misses, so detection is real but bounded.
- T1027.002partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can identify known packer signatures and heuristic anomalies, but packed executables are specifically designed to evade signature-based detection and the control's guidance acknowledges that evasive techniques like encryption can bypass these measures.
- T1027.002partialprevents — A.8.7's allowlisting (8.7a), defense-in-depth placement (8.7g), and behavioral/repair layers can prevent some packed executables from running, but the technique is specifically designed to evade the signature-based detection that is A.8.7's primary mechanism, and the clause itself states that detection-and-repair software alone is not usually adequate.
- T1027.003partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can surface steganographic payloads through signature or heuristic analysis, but the technique is designed to evade detection and the control's guidance does not guarantee it will see through every embedding.
- T1027.004partialdetects — A.8.7's scanning and allowlisting can catch some instances (known malicious source patterns, unauthorized compiler invocation via allowlisting), but T1027.004 is explicitly designed to subvert protections targeting executables/binaries by delivering text-based source code that evades signature-based malware detection, and the subsequent compilation uses legitimate system tools (csc.exe, ilasm.exe, GCC) that are not inherently malicious.
- T1027.004partialprevents — A.8.7's application allowlisting (8.7a) and website blocklisting (8.7b) can prevent some instances of T1027.004, but the technique is specifically designed to evade malware scanning by delivering source code rather than executables, and the native compilers it relies on (csc.exe, GCC, etc.) are typically permitted on systems where development occurs.
- T1027.005partialdetects — A.8.7 explicitly requires malware detection/repair software, regular automated scans of files/emails/webpages/media, collection of new malware intel, and risk-based placement (including defense-in-depth at gateways/endpoints), which surfaces the presence of modified tools whose indicators were altered to evade prior detection.
- T1027.006partialdetects — A.8.7's scanning of webpages (f.3), defense-in-depth at gateways and endpoints (g.1), and endpoint malware detection can surface some HTML smuggling instances — particularly when the constructed payload is written to disk or executed — but the technique is explicitly designed to evade the content-filtering and file-scanning mechanisms the clause centres on, since the HTML/JS carrier presents as benign text/html or text/plain and the payload is assembled client-side.
- T1027.006partialprevents — A.8.7's allowlisting (a), web/email scanning (f), and blocklisting (b) catch a slice of HTML smuggling — known-malicious HTML patterns, blocked delivery domains, and constructed payloads that fail allowlists — but the technique is purpose-built to evade content filters by constructing the payload client-side from seemingly benign MIME types, so the core mechanism is not prevented.
- T1027.006partialresponds — A.8.7's malware detection and repair software (clause f) responds to the constructed payload once HTML smuggling has executed — the deobfuscated file is scanned and removed at the endpoint — but the smuggling technique itself (bypassing content filters via Blobs/Data URLs/download attributes) is specifically designed to evade the network and gateway layers A.8.7 also relies on, so the response catches the resulting malware rather than the technique, and sophisticated payloads can still evade endpoint detection.
- T1027.007partialdetects — A.8.7's malware detection software (especially modern EDR with behavioral analysis) can detect some instances of dynamic API resolution through runtime monitoring of API call patterns and behavioral heuristics, but the technique is specifically designed to evade the primary detection mechanisms in A.8.7 (signature-based scanning of files and media), and the obfuscation itself leaves minimal static artifacts; what is caught is a slice dependent on the depth of behavioral instrumentation rather than a bounded remainder of a mechanism that mostly works.
- T1027.008partialdetects — A.8.7's malware detection software (item f) and defense-in-depth scanning (item g) can catch some stripped payloads through behavioral or heuristic analysis, but T1027.008 is specifically designed to evade the signature- and string-based detection that A.8.7's primary mechanism relies on, leaving a substantial portion of stripped payloads undetected.
- T1027.009partialdetects — A.8.7's malware detection and repair software (clause f) scans files, email attachments, downloads and webpages for malicious content, which surfaces many embedded payloads, but the technique is specifically designed to conceal malicious content within benign files and the clause itself names evasive techniques (g.2) as a challenge, so detection catches a slice rather than the bulk.
- T1027.009partialprevents — A.8.7's allowlisting (a) and scanning (f) can prevent some embedded payloads from executing, but the technique is specifically designed to evade signature-based detection by hiding within seemingly benign or trusted files, and sophisticated implementations bypass these controls.
- T1027.010partialdetects — A.8.7's primary detection mechanism is signature-based malware scanning (item f), which command obfuscation is specifically designed to evade; the clause acknowledges evasive techniques (item g.2) but does not mandate behavioral or heuristic analysis capable of catching most obfuscated commands, so detection is limited to known patterns, behavioral side-effects, and allowlisting violations rather than the obfuscation itself.
- T1027.011partialdetects — A.8.7's primary mechanism (file-based malware detection and repair software per (f)) is explicitly evaded by fileless storage, as the technique's own description notes; however, the clause's broader controls — application allowlisting (a), automated validation of system content and investigation of unauthorized amendments (d), and defence-in-depth placement (g) — can surface unauthorized registry, WMI or shared-memory writes in some implementations, so detection is a slice rather than the bulk.
- T1027.011partialprevents — A.8.7's primary mechanism (file-based malware detection and repair software) is explicitly evaded by fileless storage, which the technique description notes 'may remain undetected by antivirus and other endpoint security tools that can only access specific file formats from disk storage'; however, secondary elements — application allowlisting (a) restricting tools that write to registry/WMI/shared memory, defense-in-depth placement (g) potentially including memory or behavioral inspection, and environment isolation (k) — provide partial prevention against the storage act itself, though none of these directly target the fileless storage locations as a class.
- T1027.011partialresponds — Malware repair software can eradicate known malicious payloads stored in the Registry or shared memory, but the technique's explicit design to evade traditional file-based scanners leaves a substantial remainder of fileless artifacts unaddressed.
- T1027.012partialdetects — A.8.7's malware detection software scans email attachments, downloads and files at multiple points (f.1, f.2, f.3), and modern AV/EDR can flag suspicious LNK metadata or known icon-smuggling patterns, but the technique is explicitly designed to bypass content filters by hiding the payload reference in the icon location field rather than in the file body, so static scanning of the LNK itself often sees only benign metadata and the malicious download only occurs on execution — a slice is caught, not the bulk.
- T1027.012partialprevents — A.8.7's malware detection and repair software (f), email attachment scanning (f.2), web filtering (b), and allowlisting (a) all provide mechanisms that can prevent LNK Icon Smuggling, but the technique is specifically engineered to evade content filters by hiding the malicious URL in LNK metadata fields that standard scanners may not inspect; prevention depends on implementation depth (e.g., whether the scanner parses IconEnvironmentDataBlock), so the control reaches a slice rather than the bulk.
- T1027.012partialresponds — Malware repair software quarantines or deletes the malicious LNK file and its downloaded payloads, providing automated eradication of the technique's artifacts.
- T1027.013partialdetects — A.8.7's malware detection software (f) is exactly the layer T1027.013 is designed to defeat via obfuscation, but the clause has compensating mechanisms — (d) automated validation of software/data content investigating unapproved files, (g)(2) explicitly addressing encrypted-file delivery as an evasive technique, (a) allowlisting, and (n) staying current on new malware — that surface some obfuscated artifacts, so detection is real but bounded by the technique's core purpose of evading static signatures.
- T1027.014partialdetects — A.8.7's primary mechanism (signature-based scanning per clause f) is explicitly evaded by polymorphic code per the technique's own description, but the clause also mandates behavioural and heuristic detection through defence-in-depth placement (g), allowlisting (a), and regular content validation (d), which catch a slice of polymorphic variants whose mutation patterns or unauthorised execution surface them.
- T1027.014partialprevents — A.8.7's allowlisting (a), blocklisting (b), and defense-in-depth (g) measures can prevent some polymorphic code execution, but the technique is explicitly designed to evade the clause's primary mechanism — signature-based detection and repair software (f) — and the clause itself acknowledges that AV alone is not usually adequate, so what is prevented is a slice rather than the bulk.
- T1027.014partialresponds — A.8.7's malware detection and repair software includes a repair/eradication function that quarantines and removes malicious code once detected, which is a genuine response action; however, polymorphic code is specifically designed to evade the signature-based detection that triggers this response, so the containment and eradication only fires when heuristic or behavioral detection succeeds, leaving a bounded but real remainder.
- T1027.015partialdetects — A.8.7's malware detection software scans inside compressed archives (per f.1, f.2) and explicitly considers evasive techniques like encrypted files (g.2), so it detects compressed payloads in many cases; however, T1027.015 is specifically designed to defeat that scanning via password protection, encryption, and concatenated-ZIP tricks that some readers cannot fully parse, leaving a meaningful remainder where the technique evades detection.
- T1027.015partialprevents — A.8.7's malware scanning of files before use (f1, f2) and webpage scanning (f3) can detect and block compressed malicious payloads at the boundary, but the technique's use of concatenated archives, self-extracting archives, and compression combined with encryption specifically targets detection gaps that the control's guidance acknowledges (g2), leaving a nameable remainder.
- T1027.016partialdetects — A.8.7 mandates malware detection and repair software that scans files and data content, which can flag obfuscated malware through dynamic or behavioral analysis, but junk code insertion specifically targets static signature-based detection, leaving a gap that the control's guidance acknowledges by noting evasive techniques like encryption.
- T1027.017partialdetects — A.8.7 mandates malware detection at gateways, email servers and endpoints (clauses f.2, f.3, g.1), which can surface some SVG smuggling — known signatures, embedded `<script>` tags, or behavioural detection when the SVG executes — but the technique is specifically engineered to evade content filters by hiding payloads inside a file type treated as an image, so detection depends on inspection depth the clause does not mandate and a meaningful share of variants pass through.
- T1027.017partialprevents — A.8.7's scanning of email attachments, downloads and webpages (f), blocklisting of malicious sites (b), application allowlisting (a), and user awareness training (m) all provide mechanisms that can prevent SVG smuggling, but the technique is specifically engineered to evade these controls by hiding JavaScript inside XML-structured image files that content filters and AV engines may not parse as executable content, so prevention is real but bounded.
- T1027.017partialresponds — A.8.7's malware detection and repair software can detect and quarantine malicious SVGs that bypass initial content filters, providing containment/eradication once the technique is underway. However, the technique is specifically designed to evade content filters, and A.8.7's primary mechanisms are preventive (pre-use scanning, allowlisting, blocklisting), so the response is limited to what the malware detection can catch after delivery.
- T1027.018partialdetects — A.8.7's malware detection software (clause f) and automated validation (clause d) can flag some instances of invisible Unicode — known signatures, suspicious character distributions, or the reconstructed payload at execution — but the technique is explicitly designed to evade basic static analysis and visual inspection, which is the primary surface A.8.7's scanning mechanisms cover, so a meaningful share slips through.
- T1027.018partialprevents — A.8.7 does not prevent the insertion of invisible Unicode characters themselves, but its application allowlisting (a) and execution controls can prevent the hidden payload from being reconstructed at runtime via eval() or similar, blunting the technique's purpose; the obfuscation itself and many delivery vectors (e.g., AI prompt injection in emails, document-borne content) remain outside the clause's reach.
- T1027.018partialresponds — A.8.7 includes malware repair software that quarantines or deletes malicious files, eradicating the payload concealed by invisible Unicode, though the technique's evasive nature means it may bypass detection and avoid this response.
- T1036partialdetects — A.8.7's mechanisms — application allowlisting (a), automated validation of software/data content (d), and malware detection software (f) — do detect some masquerading by catching unauthorized software regardless of name and flagging metadata/behavior anomalies, but T1036 is specifically designed to evade these defenses, and the clause sets requirements rather than mandating the depth of behavioral or metadata analysis needed to catch sophisticated renaming and LOLBAS abuse.
- T1036partialprevents — A.8.7's allowlisting (a), automated validation of software content (d), and scanning (f) can prevent some masquerading by blocking unauthorized or altered binaries from executing, but the technique is specifically designed to evade these defenses through legitimate-looking names, metadata, and signatures, so a meaningful share of masqueraded artifacts still run.
- T1036partialresponds — A.8.7's malware repair software quarantines and eradicates malicious files, bounding the response to masquerading artifacts that are actually malware and leaving manually renamed legitimate utilities unaddressed.
- T1036.001partialdetects — A.8.7's malware detection and repair software (clause f) typically performs code signature validation as part of its scanning, and clause (g.2) explicitly calls out considering evasive techniques — invalid signatures are one such indicator — but the control does not mandate signature validation specifically, and many implementations focus on behavioral or hash-based detection that would catch the malware regardless of signature validity without surfacing the masquerade itself.
- T1036.001partialprevents — A.8.7's application allowlisting guidance (item a) would block execution of files with invalid signatures that aren't on the allowlist, and behavioral detection layers would still surface malicious activity, but the technique is specifically designed to evade signature-based trust checks which form part of the control's malware detection approach, and the clause does not mandate code signature validation as a specific control.
- T1036.002partialdetects — A.8.7 mandates scanning files and email attachments before use, which can surface the RTLO character in a filename, but the control's guidance does not require Unicode-aware filename inspection, so detection depends on implementation depth rather than being guaranteed by the clause.
- T1036.002partialprevents — A.8.7's malware scanning (f) detects the malicious payload regardless of the RTLO-disguised filename, application allowlisting (a) prevents execution of files whose true type does not match policy, and awareness training (m) helps users spot the reversed-character trick — but the RTLO character itself is a Unicode rendering issue A.8.7 does not directly address, and trained users still click, so the technique's success rate is reduced rather than eliminated.
- T1036.003partialdetects — A.8.7's allowlisting guidance (a) and behavioral components of modern malware detection can surface renamed utilities via hash or execution behavior, but the technique is specifically designed to evade the signature-based detection that is A.8.7's primary mechanism, and a renamed legitimate binary is not malware to match against signatures.
- T1036.004partialresponds — A.8.7's malware repair software can eradicate malicious tasks or services identified as malware, but this response is limited to malware and does not cover manual masquerading by an adversary using legitimate system tools.
- T1036.005partialdetects — A.8.7's allowlisting (a), content validation (d), and malware detection software (f) catch the underlying payload and flag unauthorized files regardless of their deceptive names, so the technique is detected when the malware is caught; however, the masquerading itself (the naming/location trick) is specifically designed to evade name-based scrutiny, and signature-only implementations miss novel payloads, leaving a bounded remainder.
- T1036.005partialprevents — A.8.7's allowlisting (a) and automated content validation (d) prevent the masqueraded file from executing or persisting regardless of its legitimate-sounding name, but the technique itself — placing a file with a trusted name in a trusted location — can still occur and can still partially evade name-based detection layers, so the prevention is real but bounded.
- T1036.006partialdetects — A.8.7's malware detection software scans file content and would catch the underlying malicious payload, but the space-after-filename trick itself is a UI-level user deception that malware scanning does not specifically look for; allowlisting (8.7a) and content validation (8.7d) catch a slice of unauthorized executables but not the masquerading mechanism.
- T1036.006partialprevents — A.8.7's malware detection and repair software scans files before use (f1, f2) and can flag or block a Mach-O executable disguised with a trailing space, but the technique exploits OS-level file-type determination that may bypass signature-based detection, leaving a nameable remainder.
- T1036.006partialresponds — Malware repair software quarantines or deletes identified malicious files, eradicating the payload executed via this masquerading technique.
- T1036.007partialdetects — A.8.7's malware detection and repair software (clause f) scans email attachments and downloads before use, and would detect the malicious payload behind a double extension based on content/signatures; clause d's validation of software and data content can also surface files with mismatched or deceptive extensions, and some AV heuristics specifically flag double extensions. The remainder is files that evade scanning (e.g., delivered through channels not covered by the configured scan points, or payloads using encryption/packing that delay detection until execution).
- T1036.008partialdetects — A.8.7 explicitly requires regular automated validation of software/data content, scanning of received files/email/webpages/storage media before use, and placement of detection tools based on risk (including at gateways/endpoints), which surfaces masqueraded file types via signature/extension/content mismatches; this is a genuine but minority slice of the technique (remainder includes polyglots, post-compromise in-memory use, and evasion of scanners via encryption/obfuscation).
- T1036.008partialprevents — A.8.7's malware detection and repair software scans files before use and can identify malicious payloads regardless of their disguised extension or header, but the technique's polyglot files and novel masquerading methods that evade signature-based detection leave a nameable remainder.
- T1036.009partialdetects — A.8.7 mandates malware detection software with defense-in-depth placement, and modern EDR implementations can detect orphan processes (PPID 1 adoption) and double-fork patterns, but the technique is explicitly designed to evade process-tree-based analysis and A.8.7 does not require the specific behavioral telemetry needed to catch it — implementations relying on parent-child relationships alone will miss it.
- T1036.009partialresponds — A.8.7's malware repair software can kill the malicious process and quarantine the executable, eradicating the malware that broke the process tree, though the technique's evasive nature means the response is not guaranteed.
- T1036.011partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, but overwriting process arguments is an in-memory runtime technique that leaves no file or network artifact for signature-based or file-integrity scanners to observe, so detection depends on whether the organization's implementation includes host-based behavioural monitoring or EDR, which the clause permits but does not require.
- T1036.011partialresponds — Malware repair software can terminate and quarantine malicious processes that employ argument overwriting, eradicating the technique's execution, though this is limited to activity the software classifies as malware.
- T1037partialdetects — A.8.7's malware detection software (item f), application allowlisting (item a), and regular validation of software content (item d) can surface malicious or unauthorized boot/logon scripts, but the technique specifically abuses legitimate system mechanisms where signature-based detection often misses novel persistence content, and the clause itself states that detection software alone is not usually adequate.
- T1037partialprevents — A.8.7's allowlisting (a) and malware detection (f) can catch some unauthorized or malicious script content, but T1037 specifically abuses legitimate boot/logon initialization mechanisms that must be permitted for the system to function, so the control reduces but cannot eliminate the attack surface.
- T1037partialresponds — Malware repair software can quarantine or remove identified malicious scripts used for initialization, providing partial eradication, though it may not automatically remediate adversary modifications to legitimate system scripts.
- T1037.001partialdetects — A.8.7's malware detection software can flag malicious scripts at execution, and item (d)'s automated validation of software/data content with investigation of unauthorized amendments can surface registry changes to UserInitMprLogonScript, but the technique's persistence mechanism (registry key pointing to a script) is not the primary focus of malware protection controls and many logon-script abuses use legitimate scripting tools that evade signature-based detection.
- T1037.001partialprevents — A.8.7's allowlisting (a) and automated validation of software/data content investigating unapproved files or unauthorized amendments (d) can prevent or detect unauthorized logon script additions, but the control is scoped to malware rather than abuse of legitimate Windows persistence mechanisms, and the registry-key modification central to T1037.001 is not its primary target.
- T1037.002partialdetects — A.8.7(d) requires investigating unauthorized amendments to system files and (f) requires malware detection software that could surface the malicious script payload, but the login hook is a persistence mechanism involving a specific plist modification that malware protection controls do not comprehensively monitor — detection depends on whether file-integrity-style validation and script-content scanning are actually implemented for this macOS-specific path.
- T1037.002partialprevents — A.8.7's allowlisting (a) and unauthorized-amendment detection (d) can block or surface the malicious script the hook would execute and the plist modification itself, but the technique is a persistence mechanism rather than malware delivery, and admin-level access plus legitimate-looking plist edits leave a meaningful remainder the clause does not reach.
- T1037.002partialresponds — Malware repair software can quarantine the malicious script executed by the login hook, eradicating the payload, though it may not revert the plist configuration change itself.
- T1037.003partialdetects — A.8.7 detects some instances of malicious logon scripts via malware detection software scanning script content (item f), application allowlisting detecting unauthorized software (item a), and validation of unauthorized amendments to systems (item d), but the primary detection of logon script abuse as a persistence mechanism belongs to monitoring (A.8.16) and change management (A.8.32), not malware protection.
- T1037.003partialresponds — Malware repair software can quarantine or delete a malicious logon script file if it is identified as malware, eradicating the payload, but it does not revert the Active Directory or GPO configuration changes that established the persistence.
- T1037.004partialdetects — A.8.7(d) explicitly covers investigating 'unauthorized amendments' to system software and data content, and AV scanning plus allowlisting (a) can surface malicious additions to RC scripts, but the technique modifies legitimate system startup files rather than dropping known malware, so detection depends on whether file-integrity monitoring and script-content scanning are actually configured — a slice the implementer chooses, not a bounded remainder of a mechanism that mostly works.
- T1037.004partialprevents — A.8.7's allowlisting (a), file/content validation (d), and malware detection (f) can prevent or constrain some RC script persistence attempts, but the technique requires root privileges to modify system startup files, and A.8.7 does not mandate the access controls, file integrity protections, or write restrictions that would be needed to stop a privileged actor from editing rc.local or similar files; what is prevented is a slice, not the bulk.
- T1037.004partialresponds — Malware repair software can eradicate the malicious binaries executed by the RC scripts and may clean known malicious script modifications, but manual adversary edits or non-malware persistence mechanisms remain unaffected.
- T1037.005partialdetects — A.8.7 mandates malware detection and repair software that scans files and data content, which would surface a malicious StartupItems directory during regular automated validation or on-access scanning, but the control's detection scope is risk-based and an adversary placing a legitimate-looking plist in a deprecated but still-functional path may evade signature-based detection, leaving a nameable remainder.
- T1037.005partialprevents — A.8.7's allowlisting (a) and regular validation of software/data content (d) are genuine preventive mechanisms against unauthorized executables placed in startup locations, but the control is general malware protection rather than specifically targeting persistence mechanisms, the technique is macOS-specific and uses a deprecated folder that may not exist on modern systems, and implementation coverage of startup-item directories varies.
- T1037.005partialresponds — Anti-malware software can quarantine and remove known malicious executables placed in the StartupItems directory, but will not eradicate custom or benign-looking shell scripts used for persistence.
- T1041partialresponds — A.8.7's malware repair capabilities can quarantine the malicious process driving the C2 channel, halting the exfiltration underway, though this only applies to malware-based C2 and cannot retrieve data already lost.
- T1047partialdetects — A.8.7's malware detection software (clause f) and defense-in-depth placement (clause g) can surface known malicious payloads delivered via WMI and suspicious WMI-provider child processes (e.g., wmiprvse.exe spawning scripting interpreters), but WMI is a legitimate built-in administration feature, so the technique is abuse of a trusted tool rather than malware in the traditional sense; allowlisting (clause a) does not reach it, and detection of fileless or living-off-the-land WMI usage depends on telemetry depth the clause does not mandate.
- T1047partialprevents — A.8.7's allowlisting (a) can block specific WMI client executables like wmic.exe, and malware detection (f) can catch payloads delivered via WMI, but WMI is a built-in Windows service the OS itself requires and the technique can be invoked through PowerShell, COM APIs, VBScript, .NET and other paths that cannot be removed without breaking Windows — so what is prevented is a slice, not the technique.
- T1047partialresponds — Malware repair and isolation tools can kill malicious child processes or quarantine payloads executed via WMI, but they do not typically remediate fileless WMI persistence mechanisms such as malicious event subscriptions.
- T1053partialdetects — A.8.7's malware detection software (clause f) and regular automated validation of software/data content (clause d) can surface some malicious scheduled task creation — EDR/AV products flag suspicious schtasks.exe or cron activity, and unapproved task entries are an 'unauthorized amendment' the validation step investigates. But T1053 is a living-off-the-land technique that abuses legitimate system utilities in ways that are hard to distinguish from normal administrative use, and the clause's primary mechanism (signature-based malware detection) is poorly suited to fileless or built-in-tool abuse, so detection is partial rather than mostly.
- T1053.002partialprevents — A.8.7's application allowlisting guidance (item a) can block `at.exe` execution where it is not authorized, but the technique has multiple vectors — WMI's Win32_ScheduledJob class, schtasks, and on Linux/macOS `at` is a commonly needed system utility — so the control prevents a slice of the technique's surface rather than the bulk.
- T1053.003partialdetects — A.8.7's malware detection software (item f) catches the malicious payload that cron executes, and items (a) and (d) — unauthorized-software controls and automated validation of software/data content — can surface unauthorized crontab entries; however, A.8.7 is scoped to malware protection rather than persistence-mechanism monitoring, so detection of the cron scheduling act itself depends on whether the EDR/AV implementation specifically watches crontab modifications, leaving a bounded remainder.
- T1053.005partialdetects — A.8.7's malware detection software can identify the malicious payload that gets executed via the scheduled task, and its allowlisting/unauthorized-software detection (clause a) can flag unapproved scheduled task entries, but the control's scope is malware files and their delivery rather than abuse of legitimate system utilities, so the scheduled-task mechanism itself is largely outside its detection surface.
- T1053.005partialprevents — A.8.7's allowlisting, access controls and change management can restrict who creates scheduled tasks and what payloads get scheduled, but the Task Scheduler is a required Windows component that cannot be removed and adversaries with sufficient privileges (including SYSTEM) can still register tasks — the same shape as cm-7 vs T1059, where the platform-required mechanism stays reachable and what is prevented is a slice.
- T1053.005partialresponds — Malware repair software can eradicate scheduled tasks that are identified as artifacts of a known malware infection, but this leaves tasks created for living-off-the-land execution or manual persistence unaddressed.
- T1053.006partialdetects — A.8.7's malware detection and repair software can surface the malicious payload executed by the systemd-managed service (scanning per (f)), and clause (d)'s automated validation of system content can flag unauthorized .timer/.service files in /etc/systemd/system or ~/.config/systemd/user, but the timer mechanism itself is legitimate Linux functionality that signature-based AV does not target, and the clause's primary mechanism is malware scanning rather than behavioral or file-integrity monitoring of persistence infrastructure, so detection is a slice rather than a bounded remainder of a mechanism that mostly works.
- T1053.006partialresponds — Malware repair software can quarantine the malicious payload executed by the timer, providing eradication, but this is partial because it may leave the timer configuration itself intact.
- T1053.007partialdetects — A.8.7's malware detection and repair software could surface malicious payloads within container images or running containers if scanning extends to container environments, and application allowlisting could flag unauthorized container deployments, but the control's guidance is scoped to traditional endpoints, email, and web — not container orchestration platforms — and the core mechanism of T1053.007 (abusing legitimate CronJob scheduling) is not directly addressed.
- T1055partialprevents — A.8.7's allowlisting (a), vulnerability reduction (c), and file/email scanning (f) can prevent some process injection vectors — particularly where a malicious payload is delivered as a file or where unauthorized software initiates the injection — but the technique fundamentally abuses legitimate OS APIs (CreateRemoteThread, etc.) and can be executed by already-trusted processes, so the control cannot prevent the technique as a whole.
- T1055partialresponds — A.8.7 includes response elements — the 'repair' function of malware detection software (f) and environment isolation (k) — that act on malware once detected, but these are secondary to its primary prevention/detection focus, and process injection is specifically designed to evade process-based defenses, limiting how completely these response measures engage against T1055.
- T1055.001partialprevents — A.8.7's allowlisting (a), defense-in-depth placement (g), and behavioral malware detection can prevent some DLL injection instances, but the technique is explicitly designed to evade process-based defenses and the malware-detection software A.8.7 mandates — reflective loading and module stomping use legitimate DLLs or never touch disk, so the control's primary mechanism is the very thing the technique bypasses.
- T1055.001partialresponds — A.8.7's malware detection and repair software (f) and environment isolation (k) perform some response actions — removing a detected injected DLL and quarantining affected systems — but the control is primarily preventive/detective and does not carry the full incident response process (assessment, evidence collection, eradication of the actor's foothold), so the response is a slice of the verb's denominator.
- T1055.002partialprevents — A.8.7's application allowlisting (8.7a) and prevention of unauthorized software can block the initial injector from running, but PE injection operates at the process-memory level via APIs like VirtualAllocEx/WriteProcessMemory and is explicitly designed to evade process-based defenses, which is outside A.8.7's primary scope of file scanning and software allowlisting.
- T1055.002partialresponds — A.8.7's malware detection and repair software (clause f) can remove or quarantine injected PE code once detected, and clause (k) on isolating environments provides containment, so a real response action exists; however, PE injection is specifically designed to evade process-based defenses by running in the address space of a legitimate process, and A.8.7 is primarily a preventive/detective control rather than a dedicated incident-response function, so the response reach is bounded.
- T1055.003partialprevents — A.8.7's file scanning, application allowlisting, and email/attachment checks can prevent the initial payload that performs thread hijacking from executing, but the technique itself operates in-memory on already-running legitimate processes via API calls (OpenThread, SuspendThread, WriteProcessMemory) and is explicitly designed to evade process-based defenses, so the technique's core mechanism is not stopped by A.8.7.
- T1055.003partialresponds — A.8.7's malware detection and repair software (clause f) can detect the injected code and attempt to remove it once the technique has run, and clause k's environment isolation can contain spread, but the control's primary purpose is prevention and the 'repair' software is oriented toward infected files rather than in-memory thread hijacking, which is specifically designed to evade process-based defenses.
- T1055.004partialprevents — A.8.7 prevents enabling conditions for APC injection — malware detection can catch the malicious DLL before it lands on disk, and application allowlisting (8.7a) can block the unauthorized executable that performs the injection — but the technique's core mechanism uses legitimate Windows APIs (OpenThread, QueueUserAPC) that cannot be removed, so once an attacker has code execution and a payload, the injection itself runs unimpeded, matching the cm-7 anchor's shape where platform-required functionality stays reachable.
- T1055.004partialresponds — A.8.7's malware detection and repair software can detect and quarantine the malicious DLL loaded via QueueUserAPC and isolate affected environments, but APC injection is specifically designed to evade process-based defenses by executing within a legitimate process, and the control is primarily preventive rather than responsive, so containment and eradication reach only a slice of instances.
- T1055.005partialdetects — A.8.7's malware detection and repair software (clause f) and defense-in-depth placement (g.1) can surface TLS callback injection through behavioral and memory analysis, but the technique is explicitly designed to evade process-based defenses by running under a legitimate process, and the clause sets a requirement rather than mandating the specific instrumentation depth needed to reliably catch in-memory pointer manipulation.
- T1055.005partialprevents — A.8.7's application allowlisting (8.7a), email/attachment scanning (8.7f.2), and web scanning (8.7f.3) can prevent the initial malware from reaching or executing on the system, but TLS callback injection is specifically designed to evade process-based defenses by running within legitimate process context, and the technique's own description notes it 'may evade detection from security products' — so the prevention is real but bounded by the technique's evasion design and its operation inside already-trusted processes.
- T1055.008partialdetects — A.8.7's malware detection and repair software (clause f) and defence-in-depth placement (clause g) can surface ptrace-based injection through behavioural monitoring of syscalls and process attachments, but the technique is explicitly designed to evade process-based defenses, ptrace is a legitimate debugging tool requiring behavioural context to flag, and Linux EDR coverage is less mature than Windows — so detection catches a slice rather than the bulk.
- T1055.008partialprevents — A.8.7's application allowlisting (a) can block unauthorized binaries that invoke ptrace, and its malware detection (f) can catch injected shellcode, but ptrace is a legitimate OS-provided system call the platform requires for debugging and cannot be removed without breaking the system — the technique explicitly notes evasion of security products, and the attacker can use already-allowed binaries, so what is prevented is a slice rather than the bulk.
- T1055.008partialresponds — Malware protection and repair software can terminate processes exhibiting malicious ptrace injection behaviors, providing containment and eradication for the subset of injections recognized as malicious.
- T1055.009partialdetects — A.8.7's malware detection and repair software (clause f) and defense-in-depth placement (clause g) can include EDR with process or syscall telemetry capable of surfacing /proc memory writes, but the clause does not mandate that instrumentation depth; a baseline AV implementation fully conformant with A.8.7 sees nothing of an in-process injection, and the technique is explicitly designed to evade process-based defenses.
- T1055.009partialprevents — A.8.7's allowlisting (a) and vulnerability reduction (c) can prevent some instances by restricting unauthorized tools and closing exploit paths, but T1055.009 uses legitimate Linux utilities (dd) and the /proc filesystem itself, which cannot be removed without breaking the platform, and the technique is explicitly designed to evade detection by security products.
- T1055.009partialresponds — Malware repair software can terminate malicious processes and quarantine associated files, providing containment and eradication of the injection activity, though it may not fully remediate fileless or highly evasive injections.
- T1055.011partialdetects — A.8.7's malware detection and repair software (clause f) is specifically aimed at detecting malicious code execution and can surface EWM injection through behavioral analysis, memory scanning, and API monitoring at endpoints and gateways; however, the technique is explicitly designed to evade detection by avoiding monitored APIs (WriteProcessMemory, CreateRemoteThread), running under legitimate processes, and using a small legitimate memory region, so detection is incomplete and depends on implementation depth rather than being guaranteed by the clause.
- T1055.011partialprevents — A.8.7's malware detection software and defense-in-depth placement can catch some EWM injection payloads via signatures or behavior, and application allowlisting constrains what runs, but the technique is specifically designed to evade process-based defenses by using legitimate Windows GUI APIs (GetWindowLong/SetWindowLong), avoiding monitored APIs like WriteProcessMemory/CreateRemoteThread, and executing under a legitimate process — so what is prevented is a slice, not the bulk.
- T1055.011partialresponds — A.8.7's detection and repair software can respond to EWM injection once underway through behavioral analysis, signature matching, and quarantine of the affected process, but the technique is specifically designed to evade process-based defenses by avoiding monitored APIs (WriteProcessMemory, CreateRemoteThread), using legitimate Windows APIs (SetWindowLong), and executing within legitimate processes, which limits how much of the technique the repair software can actually reach and eradicate.
- T1055.012partialdetects — A.8.7's malware detection and repair software (especially modern EDR) can surface process hollowing via behavioral indicators like suspended-process creation and memory-unmap calls, but the technique is explicitly designed to evade process-based defenses and the clause itself states detection software alone is not adequate, leaving a named remainder of hollowing variants that bypass signature and basic behavioral checks.
- T1055.012partialprevents — A.8.7's layered mechanisms — application allowlisting (a), behavioral detection in modern EDR (f/g), and defense in depth — can prevent some process hollowing instances, but the technique is specifically engineered to evade process-based defenses by running malicious code under a legitimate process, and the clause's own guidance (g)(2) acknowledges attackers' evasive techniques, so prevention is a slice rather than a closure.
- T1055.012partialresponds — A.8.7 includes response elements — malware detection and repair software can quarantine/remove detected malicious code (eradication) and isolating environments (clause k) is containment — but the control's primary purpose is prevention and detection, and process hollowing is specifically designed to evade process-based defenses, so the response capability is a secondary slice rather than the control's main function.
- T1055.013partialdetects — A.8.7's malware detection software and defense-in-depth scanning can surface some Process Doppelgänging via behavioral heuristics and memory analysis, but the technique is specifically engineered to evade the file-scanning and API-monitoring the clause primarily relies on — the malicious payload is rolled back from the filesystem and avoids monitored APIs like NtUnmapViewOfSection, VirtualProtectEx, and SetThreadContext, and the clause itself states detection software alone is 'not usually adequate.'
- T1055.013partialprevents — A.8.7's application allowlisting (8.7a), defense-in-depth placement (8.7g), and environment isolation (8.7k) can prevent some instances by restricting which legitimate processes can be spawned or loaded with modified sections, but Process Doppelgänging is specifically engineered to evade these controls: the malicious payload never persists on disk (defeating file scanning and AV), the host process is legitimate (defeating allowlisting), and TxF avoids the highly-monitored APIs that behavioral detection watches.
- T1055.014partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, but VDSO hijacking injects code into a running process's memory without writing to disk, so detection depends on whether the deployed tools include in-memory or behavioral scanning, which the clause recommends but does not require.
- T1055.014partialprevents — A.8.7's application allowlisting (a), vulnerability management (c), and blocklisting (b) can prevent the initial malicious binary or delivery vector from running, but VDSO hijacking is specifically designed to evade process-based defenses by executing within the address space of an already-running legitimate process; once a process is running and reachable, the in-memory code injection the technique performs is outside what allowlisting and file-scanning mechanisms constrain, so what is prevented is a slice of the attack chain rather than the technique itself.
- T1055.015partialdetects — A.8.7's malware detection and repair software, defense-in-depth placement, and unauthorized-software detection can surface some ListPlanting indicators (anomalous window message patterns, payload behavior, memory allocation in some implementations), but the technique is specifically engineered to evade process-based defenses by using 2-byte PostMessage/SendMessage writes instead of the highly monitored WriteProcessMemory, executing under a legitimate process via a SysListView32 callback — so what is caught is a slice rather than the bulk of the mechanism.
- T1055.015partialprevents — A.8.7's vulnerability management (c) and malware detection (f) can prevent some ListPlanting instances by patching the underlying weaknesses and catching payloads, but the technique uses legitimate Windows APIs (FindWindow, PostMessage, SendMessage) against legitimate processes with list-view controls, which cannot be blocked without breaking core Windows functionality, and the payload delivery is specifically designed to evade process-based defenses.
- T1056partialdetects — A.8.7's malware detection and repair software, application allowlisting, and scanning controls detect the malware-based input capture sub-techniques (keylogging, credential API hooking) by surfacing known keyloggers, hooking toolkits, and unauthorized software, but T1056 also includes non-malware variants such as Web Portal Capture and GUI Input Capture that A.8.7's mechanisms do not observe, and sophisticated or fileless credential API hooking can evade signature-based detection.
- T1056partialprevents — A.8.7's allowlisting (a), blocklisting (b), and malware detection (f) prevent a meaningful portion of input-capture malware — particularly traditional keyloggers — but T1056 spans variants that evade these controls: web portal capture (T1056.003) is essentially phishing rather than malware, credential API hooking (T1056.004) can abuse legitimate processes, and sophisticated loaders bypass signature-based detection, leaving a named remainder the clause does not reach.
- T1056partialresponds — Malware repair software can quarantine and eradicate software-based keyloggers and API hooks once they are running, but cannot respond to hardware keyloggers or external web portal captures.
- T1056.001partialprevents — A.8.7's malware detection, application allowlisting, and defense-in-depth measures prevent a meaningful portion of software-based keyloggers, but the technique has specific vectors — hardware keyloggers reading raw keystroke data from the hardware buffer, custom drivers, and API hooking by sophisticated malware — that fall outside or evade general malware protection, leaving a named remainder.
- T1056.001partialresponds — Anti-malware software includes repair capabilities that quarantine or remove malicious files and processes, eradicating software keyloggers that are identified as malware.
- T1056.002partialdetects — A.8.7's malware detection software and application allowlisting can catch the underlying malicious script (AppleScript, PowerShell, shell) that displays the fake dialog, but T1056.002 specifically leverages legitimate OS GUI APIs to render the prompt, so the dialog itself is indistinguishable from a real one and detection is limited to the code wrapper rather than the technique.
- T1056.002partialprevents — A.8.7's allowlisting (a), malware detection (f), and user awareness (m) prevent some GUI Input Capture instances—particularly those delivered via unauthorized scripts or known malicious payloads—but the technique can abuse legitimate system tools (PowerShell, AppleScript, shell) that allowlisting cannot remove, and its core mechanism is social engineering via fake prompts that training only partially mitigates.
- T1056.003partialdetects — A.8.7's items (a) detecting unauthorized software and (d) automated validation of software content / investigating unapproved files or unauthorized amendments directly address the unauthorized code installed on the web portal, but the technique is specifically credential-capture code on an externally facing service where detection depends on file-integrity monitoring depth and whether the script is recognized as malicious by installed AV — a slice rather than a bounded remainder of a working mechanism.
- T1056.003partialresponds — Malware repair software can quarantine or remove malicious scripts or files installed on the web portal, eradicating the credential capture mechanism, though it may struggle to cleanly restore legitimate files that were modified.
- T1059partialdetects — Preventing or detecting unauthorized scripts and interpreters through allow-listing and regular integrity validation limits adversaries’ ability to abuse command and scripting interpreters.
- T1059partialprevents — A.8.7 prevents a slice of T1059 through application allowlisting (8.7a), blocklisting of malicious sites (8.7b), email/attachment scanning (8.7f.2), and web scanning (8.7f.3), but cannot prevent the bulk of the technique because built-in interpreters (PowerShell, cmd.exe, bash) are required for system operation and cannot be removed — the same shape the CM-7 anchor teaches for this technique.
- T1059partialresponds — Malware repair software can terminate malicious interpreter processes and quarantine script files, but this containment only applies when the script or command is recognized as malware, leaving interactive or unrecognized abuse of legitimate interpreters unaddressed.
- T1059.001partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, which can detect malicious PowerShell scripts delivered as payloads, but it does not mandate the host-based monitoring depth required to observe PowerShell execution in memory or via .NET interfaces, leaving a significant portion of the technique's execution surface unobserved.
- T1059.001partialprevents — A.8.7's allowlisting guidance (a) can block powershell.exe from running, but PowerShell is a built-in system component that typically must remain available, and the technique explicitly bypasses powershell.exe via .NET's System.Management.Automation assembly — so what is prevented is a slice of execution paths, not the technique.
- T1059.001partialresponds — Malware protection and repair software can terminate malicious PowerShell processes and quarantine associated scripts, but will not contain living-off-the-land abuse that evades signature or behavioral detection.
- T1059.002partialdetects — A.8.7's malware detection and repair software (clause f) can scan for malicious content within AppleScript files and detect known-bad scripts via signatures or behavioral indicators, but AppleScript is a legitimate built-in macOS interpreter that cannot itself be flagged as malicious, so detection is content-dependent and bounded by what signatures and heuristics cover — leaving novel or lightly-obfuscated AppleScript abuse undetected.
- T1059.002partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block some AppleScript execution vectors (e.g., unauthorized osascript invocations, malicious .scpt files), but AppleScript is a built-in macOS interpreter required by the platform itself, with execution paths through Mail rules, Calendar alarms, Automator workflows, and NSAppleScript/OSAScript APIs from within legitimate binaries that the control cannot reach — the same shape as the cm-7 vs T1059 anchor, where the interpreters the platform requires cannot be removed.
- T1059.002partialresponds — Anti-malware tools can quarantine malicious AppleScript files and terminate their processes, but this eradication does not reach interactive or fileless abuse of the native scripting environment.
- T1059.003partialdetects — A.8.7's malware detection software can catch malicious payloads executed via cmd.exe, and its application allowlisting guidance (8.7a) can detect unauthorized software use, but the clause's primary mechanism is scanning for malware files rather than detecting abuse of a legitimate, built-in system tool — the Windows command shell is not malware, and behavioral detection of its malicious use is a bounded slice of what A.8.7 mandates.
- T1059.003partialprevents — A.8.7's application allowlisting (guidance a) can prevent cmd.exe from executing, but the control also permits authorized software and the Windows command shell is often required for legitimate administration, so prevention is a slice rather than a closure.
- T1059.004partialdetects — A.8.7's primary mechanism (malware detection and repair software) scans for malicious files and does not detect abuse of legitimate Unix shells, but secondary clauses — application allowlisting (a), automated validation of software content (d), and scanning of scripts — can detect unauthorized shell binaries or malicious shell scripts, giving partial coverage of the technique's execution surface.
- T1059.004partialprevents — A.8.7's application allowlisting (a) can block unauthorized shells from executing, but Unix shells are a core OS component that cannot be removed without breaking the system, so prevention is limited to restricting which shells run rather than eliminating the technique entirely.
- T1059.004partialresponds — Malware repair software can quarantine or delete malicious shell scripts, eradicating that specific payload, but this does not contain or eradicate interactive shell commands or the abuse of legitimate scripts.
- T1059.005partialprevents — A.8.7's allowlisting (a) and blocklisting (b) mechanisms can prevent unauthorized VB/VBA/VBScript execution, and email/attachment scanning (f.2) can stop malicious VB payloads before they run, but VB is built into Windows and Office as a required platform component, so it cannot be removed the way cm-7's least-functionality clause aims to do, leaving the bulk of the technique reachable.
- T1059.005partialresponds — Malware repair software can quarantine malicious Visual Basic scripts and terminate their processes, but relies on signatures or heuristics that may miss obfuscated or living-off-the-land usage of this legitimate interpreter.
- T1059.006partialdetects — A.8.7's malware detection and repair software (item f) and scanning at gateways/endpoints can detect malicious Python scripts and behaviors — known malware payloads, suspicious script content, and some behavioral indicators — but Python is a legitimate interpreter commonly present on systems, and adversaries routinely abuse it via fileless execution, obfuscation, or living-off-the-land techniques that evade signature and basic behavioral detection, leaving a substantial remainder undetected.
- T1059.006partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block unauthorized Python scripts and catch malicious ones before execution, but Python is often a required system interpreter (especially on Linux/macOS) that cannot be removed without breaking the platform, and adversaries can use compiled Python binaries or other interpreters — the same shape as the CM-7 vs T1059 anchor.
- T1059.007partialprevents — A.8.7's malware detection (scanning attachments, downloads, webpages), blocklisting of malicious sites, and application allowlisting catch a meaningful slice of malicious JavaScript delivery, but JavaScript interpreters are deeply integrated into browsers, macOS (JXA), and Windows (JScript) and cannot be removed without breaking legitimate functionality, so the technique's core execution surface remains reachable.
- T1059.007partialresponds — A.8.7's 'detection and repair software' performs an automated eradication action (quarantine/removal of the malicious JavaScript artifact) once the technique is observed, which is a response act; however, this is limited to artifact-level cleanup and does not contain the execution that already ran, investigate the foothold, or coordinate broader incident response, so the verb holds only partially.
- T1059.010partialprevents — A.8.7's allowlisting (item a), attachment/download scanning (f.1, f.2), and blocklisting (b) can prevent unauthorized AHK/AutoIT interpreters and malicious script delivery, but the technique abuses legitimate scripting tools that organizations commonly permit, and compiled .exe payloads evade script-specific controls — the same shape as the cm-7 vs T1059 anchor where interpreters the platform requires cannot be removed.
- T1059.011partialdetects — A.8.7's malware detection and repair software (AV/EDR) detects malicious Lua scripts and suspicious Lua interpreter activity through signature and behavioral detection, and the clause explicitly mandates scanning of files and monitoring for malware; the bounded remainder is that legitimate Lua interpreters being abused with novel or fileless payloads, and Lua embedded in applications, evade signature-based detection.
- T1059.011partialprevents — A.8.7's application allowlisting (a), blocklisting of malicious sites (b), file-source protections (e), and malware detection (f) prevent a slice of Lua-based attacks — those involving unauthorized interpreters, malicious downloads, or detected malicious scripts — but Lua is frequently embedded in legitimate applications and required by platforms, so an attacker using a legitimate interpreter with non-malicious-looking code is not stopped by this control.
- T1059.011partialresponds — Malware repair software can quarantine or delete malicious Lua scripts once identified, providing eradication for file-based execution, though it may not contain interactive or in-memory Lua commands.
- T1059.013partialdetects — A.8.7's primary mechanism (malware detection software per (f)) will not flag legitimate container CLI binaries like docker or kubectl, but the clause's broader scope — allowlisting of unauthorized software (a), blocklisting of malicious container registries (b), and defence-in-depth placement of detection tools (g) — can partially surface abuse of these tools through process, network and configuration telemetry, leaving the bulk of in-container command execution unobserved by malware-focused detection.
- T1068partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, which can surface known malware used for privilege escalation, but the control's scanning scope does not guarantee detection of zero-day exploits or fileless in-memory attacks that leave no scannable artifact.
- T1068partialprevents — A.8.7's vulnerability-management component (c) and allowlisting (a) reduce the exploitable surface T1068 relies on, but the control is framed around malware rather than exploitation of legitimate software; zero-days, signed vulnerable drivers (BYOVD), and kernel/service flaws in allowed software all bypass it, and A.8.8/A.8.19 carry the primary patching burden.
- T1070.006partialdetects — A.8.7's guidance includes (d) on investigating unauthorized amendments to systems and (f)/(g) on malware detection software where EDR-class tools can surface timestomp behavior, and timestomping occurs on the organization's own estate so it is within the control's scope; however, the clause's primary mechanisms target malware itself rather than anti-forensic metadata manipulation, and detection of timestomping depends on implementation choices (whether file-integrity or behavioral telemetry is in scope) rather than being mandated by the clause.
- T1070.010partialdetects — A.8.7 mandates malware detection and repair software that scans files and storage media, which can surface relocated payloads by their content or behaviour, but the technique's purpose is to evade such detection by blending into legitimate locations, so what is caught is a slice rather than the bulk.
- T1071partialdetects — A.8.7's malware detection and repair software scans network traffic and email attachments for malware before use, which can surface C2 traffic using application-layer protocols, but the control's scope is malware-specific and does not cover protocol-level anomaly detection or non-malware C2 blending.
- T1071.001partialdetects — A.8.7 explicitly covers malware detection at web gateways (g)(1), webpage scanning (f)(3), and blocklisting of malicious websites (b), all of which can surface some malicious web-protocol traffic; however, T1071.001 is specifically designed to blend in with normal HTTPS, and malware-detection tools catch payloads and known-bad destinations rather than C2 channels that use legitimate-looking traffic, leaving a large remainder undetected.
- T1071.001partialprevents — A.8.7's blocklisting of known malicious websites (b), web scanning for malware (f.3), and gateway malware detection (g.1) prevent some T1071.001 instances — particularly C2 to known malicious infrastructure or malware-laden web payloads — but the technique is specifically designed to blend with legitimate HTTP/S traffic that A.8.7 must permit, so the bulk of the technique (C2 over normal-looking web protocols to benign-looking domains) is not prevented.
- T1071.001partialresponds — A.8.7 includes isolation of environments (k) and procedures for reporting and recovering from malware attacks (l), which are genuine containment and response actions against active C2 over web protocols; however, the clause's primary thrust is prevention and detection, and the bulk of incident response (assessment, evidence collection, full eradication) belongs to A.5.26, leaving A.8.7 with a slice of the response task set.
- T1071.002partialdetects — A.8.7's scanning of data received over networks and email/file attachments (f.1, f.2) catches malicious payloads concealed within file transfer protocol traffic, but the technique's core is C2 channel abuse using legitimate protocols (SMB, FTP, FTPS, TFTP) to embed commands — a channel abuse that malware scanning does not target, so detection covers the file-borne payload slice rather than the protocol-as-C2 mechanism.
- T1071.002partialprevents — A.8.7's scanning of data received over networks (f.1) and gateway malware detection (g.1) can prevent malware from being delivered via file transfer protocols, which would prevent the C2 channel from establishing; however, the technique itself is about C2 communication that blends with legitimate traffic, and A.8.7 does not block the protocols or the C2 use case directly once malware is already present on a system.
- T1071.003partialdetects — A.8.7's malware detection software at endpoints can detect the malware's presence and behavior, and network-gateway scanning (g.1) can surface some anomalous mail-protocol traffic, but T1071.003 is specifically designed to blend in with normal SMTP/POP3/IMAP traffic, and A.8.7's primary focus is on malware delivery and presence rather than C2-channel detection — the latter is more the remit of A.8.16 monitoring.
- T1071.004partialdetects — A.8.7's malware detection software, particularly when placed at network gateways per the defense-in-depth guidance in (g1), can detect some DNS tunneling through anomaly detection or signatures for known tunneling tools, and (g2) explicitly contemplates evasive protocol use; however, the clause's primary mechanisms (signature-based file scanning, allowlisting, email/web scanning) target malware delivery rather than C2 channels, and sophisticated DNS tunneling with low beacon frequency can evade detection.
- T1072partialdetects — A.8.7's malware detection and repair software (clause f) can catch malicious payloads pushed through deployment tools, and allowlisting (clause a) can flag unauthorized software being distributed, but the core of T1072 — abusing legitimate, authorized configuration management infrastructure (SCCM, Intune, etc.) to execute commands — is not what A.8.7 detects; the abuse of the tool itself is invisible to malware-focused scanning, and behavioral detection of the deployment tool's misuse belongs to A.8.16 monitoring rather than A.8.7.
- T1080partialdetects — A.8.7 explicitly requires regular automated validation of software/data content (d), scanning of received files/storage media/email/webpages (f), collection of new malware intel (n-o), and placement of detection tools based on risk (g) — all of which surface tainted shared content or binary infections before or on access, with the bounded remainder being novel/encrypted/obfuscated variants or infections outside scanned shares.
- T1080partialprevents — A.8.7's application allowlisting (8.7a), malware detection scanning of network-received data (8.7f.1), and defense-in-depth at gateways (8.7g.1) all bear on shared-content delivery, but the technique is designed to evade these controls by embedding payloads in otherwise valid files (.LNK, .BAT, .VBS, binary appends) on internal shares that may sit outside the scanned perimeter, and A.8.7 itself states that detection software alone is not usually adequate.
- T1080partialresponds — A.8.7 deploys malware repair software that quarantines or deletes the malicious payloads added to shared storage, containing the spread and eradicating the artifact, though it does not revoke the access the adversary used to taint the share.
- T1090partialdetects — A.8.7 mandates malware detection and repair software that scans network traffic, email, and webpages, which can surface proxy-based C2 channels where the proxy traffic traverses the organization's network or endpoints, but the control's scope is malware-focused and does not inherently detect proxy usage for non-malware purposes or encrypted proxy traffic that evades signature-based scanning.
- T1090.001partialdetects — A.8.7's malware detection software (item f) can flag known proxy tools like HTRAN, ZXProxy and ZXPortMap by signature, and application allowlisting (item a) can detect unauthorized proxy utilities, but the technique's core mechanism — redirecting C2 traffic through internal systems using built-in OS capabilities (netsh, socat, SSH tunnels) or legitimate dual-use tools — produces no malware signature and is fundamentally a network-behavior pattern outside A.8.7's malware-centric scope.
- T1090.001partialresponds — A.8.7's malware repair software can quarantine or remove the malicious tools used to create the internal proxy, but it cannot eradicate the use of repurposed legitimate system tools for port forwarding.
- T1092partialdetects — A.8.7's removable media scanning (f)(1) and detection of unapproved files (d) can surface malicious payloads transferred via removable media, but T1092 is a C2 channel that may carry legitimate-looking commands or scripts rather than detectable malware, so the C2 communication itself is largely outside the clause's detection scope.
- T1092partialprevents — A.8.7's scanning of data on electronic storage media (f.1) and protective measures for files from external media (e) can catch malicious files transferred via removable media, but the technique presupposes both hosts are already compromised and the core C2 channel between them is not what A.8.7's prevention mechanisms (allowlisting, blocklisting, vulnerability reduction) are designed to stop.
- T1092partialresponds — Malware repair software can quarantine the malicious executables facilitating the removable media communication, eradicating the adversary's mechanism, though it may not identify or remove the command files themselves.
- T1102partialdetects — A.8.7's malware detection can identify the implant/binary that uses the web service for C2 (via signatures, behavior, or allowlisting of unauthorized software), but the T1102 technique itself — C2 over legitimate HTTPS services like Google or Microsoft — is designed to blend with expected traffic and is not what signature-based malware detection looks for; the C2 channel remains invisible even when the implant is caught.
- T1102.001partialdetects — A.8.7 mandates scanning of webpages and network traffic for malware, which can surface the initial beacon to a dead-drop resolver, but the technique's use of legitimate encrypted Web services and obfuscated content limits detection to a slice rather than the bulk of instances.
- T1104partialdetects — A.8.7's network gateway malware detection (g.1) and endpoint scanning can surface C2 callbacks and RAT components associated with multi-stage channels, but the technique's defining feature — the multi-stage architecture itself — is a C2 design choice that malware-focused detection is not built to identify, so what is caught is incidental rather than the technique's core mechanism.
- T1104partialprevents — A.8.7 prevents the RATs and loaders that constitute the multi-stage C2 from executing on the endpoint (via allowlisting, scanning, and gateway controls), which prevents the technique from being established; however, the control addresses malware execution and delivery, not C2 channel architecture itself, so once malware runs the multi-stage design is unaffected.
- T1104partialresponds — A.8.7 mandates malware repair software that can quarantine or remove the remote access tools used to establish the multi-stage channels, eradicating the local component of the technique.
- T1105partialdetects — A.8.7 mandates scanning files, email attachments, and webpages for malware before use, which detects malicious tools transferred via those channels, but it does not cover transfers through encrypted protocols or legitimate services like Dropbox syncing.
- T1105partialresponds — A.8.7's malware detection and repair software (clause f) can detect and remove tools that have been transferred in, and clause (k) on isolating environments provides containment — both are genuine response actions against T1105 once the transfer has occurred. However, A.8.7 is primarily framed as a prevention control, and the response capability is bounded by what the detection software can recognize: sophisticated or evasive transferred tools slip past, and the clause does not address the broader incident-response tasks (assessment, evidence collection, closure) that A.5.26 carries.
- T1106partialdetects — A.8.7 requires malware detection and repair software to scan files, email, webpages and storage media, which surfaces malware that uses native API calls once it is written to disk or traverses a network boundary, but the control's scanning scope does not reach in-memory-only or fileless native API execution that never touches a scanned artifact.
- T1106partialresponds — Malware repair software (such as EDR) can contain the technique by terminating processes that exhibit malicious API calls, but adversaries can bypass user-mode API hooks by using direct syscalls.
- T1111partialprevents — A.8.7's malware detection and repair software, application allowlisting, and endpoint scanning can catch the keyloggers that T1111 commonly relies on to capture smart card PINs and token inputs, but the technique also succeeds through non-malware vectors (SMS interception via compromised telecom services, phishing for one-time codes, physical token capture) that A.8.7 does not address.
- T1112partialdetects — A.8.7's malware detection software (especially modern EDR-style tools under 8.7f/g) and application allowlisting (8.7a) can detect some registry modifications — particularly those tied to known malware or unauthorized software installations — but the clause's primary detection mechanisms are file, email, attachment and web scanning, not registry monitoring, so many T1112 modifications (defense-evasion tweaks, null-character hidden keys, remote registry lateral movement) fall outside its scope.
- T1112partialresponds — Malware repair software can eradicate malicious registry keys when cleaning up a known infection, but it does not remediate manual modifications made by an adversary using built-in tools.
- T1113partialdetects — A.8.7 mandates malware detection software and regular scanning of files, email, and webpages, which can surface screen-capture malware if it is delivered as a file or downloaded, but the control does not mandate host-based behavioral monitoring that would detect a native API call like CopyFromScreen executed by a legitimate process.
- T1119partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can surface automated collection tools if they are known malware, but the control's detection scope is limited to malware signatures and heuristics, not the broader behavioral indicators of automated data collection.
- T1123partialdetects — A.8.7's malware detection and repair software (item f) detects the malware component when T1123 is delivered via malicious code, but the technique explicitly also operates through legitimate voice/video applications and OS APIs that A.8.7 does not monitor, and even for malware-based cases the control detects the malicious software rather than the audio-capture behavior itself.
- T1123partialprevents — A.8.7's allowlisting, malware detection, and blocklisting prevent the malware/script-based delivery of T1123, but the technique also operates through legitimate voice/video applications and OS audio APIs that A.8.7 does not control, leaving a meaningful remainder where audio capture proceeds without malware to detect.
- T1123partialresponds — A.8.7's malware repair software can quarantine and eradicate malicious executables performing audio capture, but this provides only a partial response because it does not address capture performed via scripts or abused legitimate applications.
- T1125partialdetects — A.8.7's detection mechanism (malware detection and repair software) catches the subset of T1125 instances delivered via malware (e.g., FruitFly, Proton), but the technique is also executed through legitimate OS APIs and applications where no malware is present, and the clause itself acknowledges evasive techniques that bypass signature-based detection — so what is observed is a slice, not the bulk with a bounded remainder.
- T1125partialprevents — A.8.7's malware detection, application allowlisting, and vulnerability reduction can prevent the malware/script-based pathway of T1125, but the technique explicitly includes capture via legitimate applications (e.g., video call services) using allowed OS APIs, which A.8.7 does not address, and sophisticated malware can evade detection.
- T1125partialresponds — A.8.7's malware detection and repair software (item f) can identify and remove the malware performing video capture once it is running on a system, which is containment and eradication — the act `responds` describes; however, novel or fileless malware, techniques abusing legitimate OS APIs for camera access, and harm already realised (video already captured and exfiltrated) all sit outside what the clause reaches, so the grade is partial rather than mostly.
- T1127partialdetects — A.8.7's malware detection software (item f) can detect malicious payloads proxied through trusted developer utilities via signature and behavioral scanning, but the technique is specifically designed to evade these controls through reputation hijacking of signed utilities, and application allowlisting (item a) is bypassed by design — so detection catches a slice rather than the technique's core mechanism.
- T1127partialprevents — A.8.7's application allowlisting (a) and malware detection (f) are the primary mechanisms T1127 is explicitly designed to evade — the technique uses legitimate, signed developer utilities that are typically permitted and proxies execution through trusted processes to bypass application control. Some prevention remains through defense-in-depth placement (g), behavioral detection of the proxied payload, vulnerability reduction (c), and user awareness (m), but the technique's core mechanism is the bypass of these very controls, leaving a substantial remainder where the trusted utility runs the payload unimpeded.
- T1127partialresponds — Anti-malware tools can terminate the offending process and quarantine the malicious payload, but the trusted developer utility itself is legitimate and remains on the system.
- T1127.001partialprevents — A.8.7's allowlisting (8.7a) is the primary prevention mechanism but is specifically defeated by T1127.001 because MSBuild.exe is a signed Microsoft binary that allowlists typically permit; other A.8.7 measures (8.7c reducing attack surface, 8.7f malware detection, 8.7g defense-in-depth) can catch some instances, but MSBuild is often a required build tool that cannot be removed, and the inline-task payload is designed to evade basic AV, leaving a bounded slice of prevention rather than full coverage.
- T1127.001partialresponds — A.8.7's malware detection and repair software performs a real response action (quarantine/remove of detected malware) and clause (k) calls for isolating environments where catastrophic consequences can occur — both are containment/eradication slices. But the bulk of response — incident assessment, evidence collection, broader containment of the running technique, and closure — is performed by A.5.26, so A.8.7's contribution is a minority of the response task set, not the core.
- T1127.002partialprevents — A.8.7's allowlisting (a), blocklisting (b), and malware scanning (f) can constrain ClickOnce abuse, but ClickOnce is a built-in trusted Windows component (DFSHIM/DFSVC.EXE) that legitimate business processes depend on, so it cannot be removed wholesale; the technique also relies on user execution of a popup, which the control's mechanisms only partially address.
- T1127.002partialresponds — Malware repair software quarantines the malicious files and terminates the associated processes, containing and eradicating the malware executed via ClickOnce.
- T1127.003partialdetects — A.8.7's malware detection and repair software, allowlisting, and behavioral monitoring can surface malicious .jam script content and anomalous build-utility execution, but the technique is explicitly designed to subvert application control by proxying through a reputable developer tool, so detection is implementation-dependent and a meaningful slice of instances evade signature and Smart App Control-style checks.
- T1127.003partialprevents — A.8.7's allowlisting (a) and malware detection (f) can prevent JamPlus abuse when the tool is not on the allowlist or when the malicious .jam script content is caught, but the technique is specifically designed to subvert application control by abusing a legitimate developer tool, so prevention is a slice rather than the bulk.
- T1127.003partialresponds — Malware repair software can terminate the malicious process and quarantine the associated script, though the legitimate JamPlus utility itself remains untouched.
- T1129partialprevents — A.8.7's application allowlisting (clause a) can be configured to restrict unauthorized DLL/shared-library loading, and malware detection (clause f) can catch the payload before it is loaded, but T1129 uses legitimate OS APIs (LoadLibrary, dlopen) that cannot be blocked without breaking the system, and most allowlisting implementations cover executables rather than DLLs, leaving a substantial remainder.
- T1132.002partialresponds — A.8.7's malware repair capabilities can quarantine the malicious process and executable generating the encoded C2 traffic, eradicating the source of the technique, though it does not act on the network traffic itself.
- T1134partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, but token manipulation operates in-memory through legitimate Windows API calls without necessarily writing malicious files to disk, so detection depends on whether the organization's risk-based tool placement and configuration (clause g) includes behavioral or endpoint detection capabilities that can observe this technique.
- T1134.004partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, but parent PID spoofing is an in-memory process manipulation technique that leaves no file artifact for a signature-based scanner to observe, so detection depends on whether the organization's implementation includes behavioral or host-based monitoring beyond the clause's core file-scanning guidance.
- T1137partialdetects — A.8.7's malware detection and repair software (clause f) and automated validation of software content (clause d) can surface malicious macros in templates/add-ins and unauthorized file amendments, but T1137 specifically abuses legitimate Office features (Outlook rules, forms, Home Page, template macros) whose persistence behaviour is configuration rather than malware-bearing content, so a substantial slice of the technique's mechanisms fall outside what the detection software is designed to flag.
- T1137partialprevents — A.8.7's malware detection (f) and application allowlisting (a) can catch malicious macros and unauthorized add-ins, but T1137 abuses legitimate Office persistence mechanisms (Outlook rules, forms, Home Page) that traditional malware protection does not reliably prevent, leaving a bounded remainder.
- T1137partialresponds — Malware repair software can quarantine and eradicate file-based persistence mechanisms like malicious Office templates and add-in DLLs, but typically does not remediate configuration-based persistence such as malicious Outlook rules.
- T1137.001partialprevents — A.8.7's malware detection (f) and application allowlisting (a) can prevent some instances of malicious template macros by detecting malicious VBA in template files or blocking unauthorized modifications, but the technique leverages legitimate Office functionality, can be delivered via trusted documents or trusted locations, and sophisticated macros evade signature-based detection, leaving a meaningful remainder.
- T1137.001partialresponds — Malware repair software can quarantine or delete the malicious Office template file to eradicate the persistence mechanism, though it may not revert associated registry modifications like GlobalDotName.
- T1137.002partialdetects — A.8.7 mandates malware detection and repair software (item f) and automated validation of software/data content (item d), which can surface the malicious DLL on disk or flag the suspicious Office Test registry modification; however, the technique abuses a legitimate Office debugging feature, the registry key itself is not inherently malicious, and signature-based detection will miss novel or custom DLLs, so detection is a slice rather than a bounded remainder of a working mechanism.
- T1137.002partialprevents — A.8.7's malware detection (f) and application allowlisting (a) can catch or block the malicious DLL that the Office Test registry key loads, and change-management validation (d) can surface unauthorized registry amendments, but the Office Test key itself is a legitimate Microsoft feature that the clause does not remove, and a custom or repurposed DLL not in malware databases or allowlists can still execute, so prevention is real but not complete.
- T1137.002partialresponds — Malware repair software can quarantine the malicious DLL executed via the Office Test key, eradicating the payload, though it may not revert the abused registry key itself.
- T1137.003partialdetects — A.8.7's email/attachment scanning (f.2) and malware detection software can surface the crafted email or the code execution, but Outlook forms are a legitimate feature being abused for persistence, so signature-based detection often misses the form itself and behavioural coverage depends on implementation depth.
- T1137.003partialprevents — A.8.7's email/attachment scanning (8.7f.2), application allowlisting (8.7a), and regular validation of software/data content (8.7d) can partially prevent malicious Outlook forms from being delivered or persisting, but the technique abuses legitimate Outlook functionality rather than introducing external malware, so general malware detection and allowlisting leave a bounded remainder where the form executes through Outlook's own mechanism.
- T1137.003partialresponds — A.8.7 mandates malware detection and repair software, which can quarantine or remove malicious Outlook forms once detected, providing partial eradication of the persistence mechanism.
- T1137.004partialdetects — A.8.7's malware detection and repair software (clause f) can scan the malicious HTML payload when it is loaded or stored, and web/email scanning may catch the content, but the persistence mechanism itself — a registry/configuration change to Outlook's Home Page setting — is not what A.8.7 monitors, and the technique abuses a legitimate Outlook feature where the configuration change appears benign to malware-focused detection.
- T1137.004partialprevents — A.8.7 provides multiple defense layers relevant to this technique—malware detection (f) can scan malicious HTML payloads, blocklisting (b) can block external URLs used for Home Pages, and regular validation of software/data content (d) can surface unauthorized Outlook configuration changes—but the technique abuses a legitimate built-in Outlook feature, so application allowlisting (a) does not remove the attack surface, and novel or fileless HTML payloads can evade signature-based detection, leaving a bounded remainder the control cannot reach.
- T1137.005partialdetects — A.8.7's email attachment scanning (f.2) can catch malicious payloads in the crafted email that triggers the rule, and the clause's broader scope (a, d) covers detecting unauthorized software and unapproved changes, but the core mechanism of T1137.005 — abusing legitimate Outlook rules as a persistence vehicle — is a configuration abuse of a trusted application rather than a malware file, and traditional malware detection software is not primarily designed to surface malicious Outlook rules themselves.
- T1137.005partialprevents — A.8.7's email and attachment scanning (clause f) can catch the crafted email or its malicious payload before the Outlook rule triggers code execution, but the rule mechanism itself is a legitimate Outlook feature being abused on an already-compromised system, and A.8.7 does not address rule creation, modification, or the persistence mechanism — so prevention is limited to the payload delivery step.
- T1137.006partialdetects — A.8.7's guidance explicitly includes detecting unauthorized software via allowlisting (a) and regular automated validation investigating unapproved files or unauthorized amendments (d), both of which can surface malicious Office add-ins; however, the technique abuses legitimate Office functionality and signed/approved add-ins can still be malicious, so detection depends on allowlist scope and malware-detection heuristics rather than being inherent to the control's mechanism.
- T1137.006partialprevents — A.8.7's allowlisting (a) and malware detection (f) can prevent unauthorized or malicious Office add-ins from loading, but add-ins are a legitimate Office feature with many signed variants, so blocking is selective rather than comprehensive and detection of malicious add-ins is not guaranteed.
- T1137.006partialresponds — A.8.7's malware detection and repair software (clause f) can detect and remove a malicious Office add-in once installed, which is an eradication action against the persistence mechanism; however, this is a narrow slice of response — it does not include broader containment, evidence collection, or recovery activities (those live in A.5.26/A.5.28/A.8.13), and add-ins using evasion or living-off-the-land techniques may not be caught, so the response is real but bounded.
- T1140partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages before use, which can surface deobfuscated or decoded payloads at the point they are written to disk or executed, but the technique's deobfuscation step itself—using built-in system utilities like certutil or copy /b—occurs in memory or as a legitimate process and is not inherently malicious, so detection depends on the tool's ability to recognize the resulting artifact rather than the act of decoding.
- T1140partialprevents — A.8.7 prevents some instances of T1140 through malware detection software that can flag malicious payloads during/after decoding, application allowlisting that can restrict unauthorized tools, and scanning that catches obfuscated content at delivery — but the technique uses built-in system utilities (certutil, copy, type) that are legitimate and typically required, so the core decoding step cannot be blocked wholesale, leaving a substantial remainder.
- T1176partialdetects — A.8.7's detection mechanisms (allowlisting per (a), automated validation of software content per (d), and EDR/AV per (f)) can surface some malicious extensions, but browser and IDE extensions are a known blind spot for traditional malware detection — they run inside trusted host processes, are often installed via official marketplaces, and the technique description explicitly notes detection is challenging because extensions blend into normal workflows.
- T1176partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent unauthorized or malicious extensions from installing, but T1176 is specifically designed to operate within trusted software ecosystems (official marketplaces, legitimate-looking extensions) where these controls are often bypassed or insufficiently granular to cover extension-level permissions.
- T1176partialresponds — Malware repair software can quarantine or remove extension files identified as malicious, but this does not address the adversary's abuse of benign extensions.
- T1176.001partialdetects — A.8.7's malware detection and repair software (item f), allowlisting (item a), and automated validation of software content (item d) can surface malicious or unauthorized browser extensions, but the technique is explicitly noted as able to evade automated scanners via app-store distribution and silent preference-file manipulation, leaving a substantial remainder undetected.
- T1176.001partialprevents — A.8.7's application allowlisting (a), blocklisting (b), malware detection (f), and validation (d) can prevent some browser extension abuse, but the technique has multiple bypass vectors — installation via trusted app stores, social engineering, silent installation by an already-compromised actor modifying preference files, and the technique's own note that malicious extensions can defeat automated scanners — so the control's mechanisms cover a slice rather than the bulk of the technique's surface.
- T1176.002partialdetects — A.8.7's malware detection software (f) and automated validation of software content (d) can surface some malicious IDE extensions through signatures, behavioral indicators, or unapproved-file checks, but the technique is specifically designed to operate within trusted IDE processes where standard AV/EDR has limited visibility, and extensions distributed through official marketplaces often evade signature-based detection.
- T1176.002partialprevents — A.8.7's application allowlisting (a), scanning of files from external networks (e, f), and regular validation of software content (d) can constrain some IDE extension abuse, but the control does not specifically address in-application extensions distributed through trusted marketplaces, and developers legitimately need to install extensions, so a meaningful slice of the technique remains reachable.
- T1176.002partialresponds — Malware repair software can quarantine or remove malicious IDE extensions once identified, but this provides no response against the adversary's abuse of benign extensions (such as Remote SSH) which the technique explicitly includes.
- T1185partialdetects — A.8.7 explicitly requires malware detection/repair software with regular automated scans of systems, received data, email/IM attachments, downloads, and webpages (including at multiple points like gateways, endpoints, and servers per defence-in-depth), which surfaces browser-injected malware or malicious extensions used in T1185; the remainder is non-malware session hijacking variants that evade detection.
- T1185partialprevents — A.8.7's application allowlisting (a), vulnerability management (c), and malware detection (f) can block malicious browser extensions and known injection tooling, but T1185 frequently abuses legitimate browser functionality (extensions, proxy settings, built-in pivoting features) and inherent browser vulnerabilities that allowlisting and signature-based scanning do not catch, leaving a substantial slice of the technique reachable.
- T1185partialresponds — A.8.7's malware detection and repair software (clause f) can detect and quarantine the malicious browser extensions or injected code that enable session hijacking, and clause (l) defines procedures for reporting and recovering from malware attacks — these are genuine response actions against the malware component. However, the core of T1185 — use of already-stolen cookies/sessions to pivot through an authenticated browser into the intranet — is not contained by removing the injection mechanism, and the adversary's lateral movement through the hijacked session is outside A.8.7's reach.
- T1187partialprevents — A.8.7's email attachment scanning (8.7f.2), application allowlisting (8.7a), and external-file protective measures (8.7e) can block some T1187 vectors like spearphishing documents with template injection or .SCF/.LNK files delivered via email, but the technique is fundamentally a credential-capture/abuse of Windows SMB/WebDAV auto-authentication protocol behavior rather than malware delivery, so EfsRpcOpenFileRaw and on-share file placement vectors fall outside the control's mechanism.
- T1189partialresponds — A.8.7's malware repair software can quarantine the malicious payloads or kill processes spawned by the drive-by compromise, but does not address non-malware exploitation outcomes such as session hijacking.
- T1190partialdetects — A.8.7's malware detection and repair software can surface post-exploitation artifacts (web shells, droppers, backdoors) left after a successful exploit, and its webpage scanning (f.3) catches malicious payloads served during exploitation, but the core of T1190 — the exploitation of a vulnerability itself — is not what malware detection observes, and many T1190 instances (SQL injection, SSRF, fileless exploitation of network devices) leave no malware for A.8.7 to find.
- T1190partialprevents — A.8.7c explicitly addresses 'reducing vulnerabilities that can be exploited by malware' through technical vulnerability management, which is directly relevant to T1190's exploitation of weaknesses in public-facing applications; however, the actual vulnerability management is delegated to A.8.8/A.8.19, and A.8.7's primary mechanism (malware detection and repair) acts after exploitation has already achieved initial access, so the prevention is partial rather than complete.
- T1195partialdetects — A.8.7's detection mechanisms — malware detection software (item f), regular automated validation of software and data content including investigation of unapproved files or unauthorized amendments (item d), and allowlisting (item a) — can surface some supply chain compromises once the malicious code is present on the estate, but the clause is built for malware arriving through traditional vectors (email, web, downloads) and is not designed to verify the integrity of software received from trusted suppliers; signed or trusted-channel supply chain malware routinely evades AV, and sophisticated compromises do not produce the unauthorized-file or unauthorized-amendment signals item (d) is written to catch.
- T1195partialresponds — A.8.7's malware repair capabilities can quarantine and eradicate the malicious payloads delivered to the estate via a compromised supply chain, though it cannot address the upstream manipulation itself or non-malware hardware compromises.
- T1195.001partialdetects — A.8.7's malware detection software (clause f) scans downloads and dependencies, and application allowlisting (clause a) can flag unauthorized software, so some instances of compromised dependencies are detected; however, the technique is explicitly designed to evade signature-based scanning (e.g., typosquatted packages, malicious code in legitimate-signed updates, natural-language agent skill files that pass registry screening), and the clause itself states that malware detection software alone is not usually adequate.
- T1195.001partialprevents — A.8.7's allowlisting (a), validation of software content (d), and scanning of received data (f) provide some prevention against malicious dependencies, but the technique is specifically designed to evade these controls — malicious code embedded in trusted packages and natural-language agent skill files passes signature-based scanning and registry screening, so what is prevented is a slice rather than the bulk.
- T1195.002partialdetects — A.8.7's clause (d) requires automated validation of software content and investigation of unauthorized amendments, and (f) requires malware scanning — both can surface some supply-chain tampering post-receipt. But the technique is specifically designed to evade these mechanisms: the tampered software is signed and appears legitimate, traditional AV signatures miss novel or targeted implants, and integrity validation is scoped to critical systems rather than universally applied. The clause does not mandate code-signing verification, SBOM analysis, or update-channel authentication that would be needed for comprehensive detection.
- T1197partialdetects — A.8.7's malware detection and repair software (f) and automated validation of software/data content (d) can surface the payloads BITS downloads when they are known malware or unauthorized files, but the technique is specifically designed to evade these controls: BITS is a legitimate Windows service, jobs are self-contained in the BITS database without new files or registry modifications, and the abuse mechanism itself (BITSAdmin/PowerShell BITS cmdlets) is not what traditional AV looks for, leaving the core technique largely invisible.
- T1197partialprevents — A.8.7's allowlisting (a) can block BITSAdmin.exe and network-gateway scanning (g) can block malicious payloads fetched via BITS, but BITS is a built-in Windows service that cannot be disabled without breaking the OS, and PowerShell's BITS cmdlets remain available to create jobs, so the core abuse vector is not closed by the clause's general malware controls.
- T1197partialresponds — A.8.7's malware repair software can quarantine malicious payloads downloaded or executed via BITS, but it typically does not remove the malicious BITS jobs themselves from the system database.
- T1202partialdetects — A.8.7's malware detection software (f) and application allowlisting (a) are primarily signature- and allowlist-based, and T1202 specifically abuses legitimate signed Windows utilities (Forfiles, pcalua.exe, Scriptrunner.exe, ssh.exe) precisely to evade these controls; however, the clause's defense-in-depth guidance (g) and behavioral aspects of modern implementations can surface abuse patterns, yielding partial but not mostly detection.
- T1203partialprevents — A.8.7(c) explicitly addresses reducing vulnerabilities exploitable by malware via technical vulnerability management (delegated to 8.8/8.19), and (a)/(b) add allowlisting and malicious-site blocklisting that constrain some exploitation paths, but T1203 targets vulnerabilities in legitimate, authorized client software (browsers, Office, third-party apps) where allowlisting does not apply, zero-days and unpatched flaws remain, and many exploits require no user action — so the clause reduces the attack surface without closing it.
- T1203partialresponds — A.8.7's malware repair software can terminate exploited processes and quarantine dropped payloads if they are identified as malware, but leaves exploitation that executes non-malware payloads unaddressed.
- T1204partialprevents — Application allow-listing and pre-use malware scanning of attachments, downloads, and web content directly block or detect the user-execution vectors that adversaries rely on to launch malicious payloads.
- T1204partialdetects — A.8.7 mandates malware detection and repair software that scans files, email attachments, and webpages before use, which surfaces malicious payloads delivered for user execution, but the control's detection scope is limited to known or suspected malware signatures and cannot identify novel or obfuscated payloads that evade signature-based scanning.
- T1204partialresponds — A.8.7's malware detection and *repair* software (clause f) and isolation of environments (clause k) act on malware once it has executed — quarantining, removing and containing it — which is the shape of `responds`. But the clause is primarily framed as preventive and detective; the response contribution is the 'repair' slice and isolation, not the full incident-response procedure (assessment, evidence collection, eradication of foothold, closure) that lives in A.5.26. A slice of the response tasks, not the bulk.
- T1204.001partialdetects — A.8.7 explicitly mandates detection of malicious websites via blocklisting (b), webpage scanning when accessed (f.3), and defence-in-depth placement at gateways and endpoints (g.1), all of which surface a malicious link as it is reached; the bounded remainder is zero-day or rapidly rotated malicious links not yet on blocklists and evasion via encryption (g.2), which the clause itself names.
- T1204.001partialprevents — A.8.7's malware detection and repair software scans webpages and email attachments before use, which can block the malicious link from delivering its payload, but the user clicking the link is a human action that scanning cannot fully prevent, leaving a bounded remainder.
- T1204.001partialresponds — A.8.7's malware detection and repair software responds to the malware execution that follows the user clicking a malicious link — detecting and removing the malicious payload once it has landed and begun executing — but the response is bounded to the malware consequence rather than the user action itself, and remains partial because encrypted payloads, zero-days, and authorized exceptions (clause i) can let execution proceed before containment.
- T1204.003partialdetects — A.8.7's scanning of data received over networks (f) and automated validation of software and data content (d) could extend to container/AMI images, and threat intelligence (n) covers new malware, but the clause's examples and focus are on traditional vectors (email attachments, webpages, files) rather than cloud/container images, so a conformant implementation may not catch this technique.
- T1204.003partialprevents — A.8.7's allowlisting (a), image scanning (f), and user awareness (m) directly address malicious image deployment, but the technique is specifically designed to bypass initial-access controls by having users willingly deploy images from public repositories, and the control cannot fully prevent a legitimate operational activity that adversaries exploit.
- T1204.003partialresponds — A.8.7's malware repair software and isolation procedures can quarantine or remove the malicious payloads executing within the deployed instance, but do not eradicate the malicious image itself from the cloud environment or container registry.
- T1204.004partialdetects — A.8.7's detection mechanisms (email/attachment scanning, web scanning, endpoint malware detection) are partially effective, but T1204.004 is specifically engineered to bypass them by relying on user action — the code is typically displayed as text on a webpage or in an email body rather than delivered as a scannable attachment, and is often obfuscated to evade web scanning; residual detection may occur at execution time through endpoint telemetry, but the technique's design defeats the primary detection vectors A.8.7 mandates.
- T1204.004partialprevents — A.8.7's user awareness/training (clause m) and malicious-website blocklisting (clause b) partially address the user-driven execution that T1204.004 relies on, but the technique is explicitly designed to bypass the control's core technical mechanisms — malware detection (clause f) cannot scan code pasted directly into a Run dialog or terminal, and application allowlisting (clause a) cannot remove the built-in interpreters (PowerShell, cmd, bash) the technique requires.
- T1204.005partialprevents — A.8.7's allowlisting (a), malware scanning (f), protective measures for external software (e), and user awareness (m) all address malicious library installation, but the technique is specifically engineered to bypass these controls — packages masquerade as legitimate code, signature-based scanning misses them, allowlisting is impractical in development environments, and trained developers still install typosquatted or backdoored packages. The control reduces the rate but cannot close the vector.
- T1205partialdetects — A.8.7's malware detection software can identify the listening component (custom binaries, libpcap-based tools, known port-knockers) on disk or in memory, but the technique is specifically engineered to evade file-based detection: signaling packets resemble ordinary failed connections to closed ports, Wake-on-LAN is a legitimate hardware feature, and the malware often uses raw sockets or libpcap without leaving a signatureable artifact — so the signaling behavior itself is outside A.8.7's scanning scope.
- T1205partialprevents — A.8.7's allowlisting, scanning of downloads/email, and blocklisting can prevent the malware that performs traffic signaling from being installed in the first place, but the signaling mechanism itself cannot be stopped by malware controls once the implant is running, and the clause explicitly states that detection-and-repair software alone is not usually adequate.
- T1205.001partialprevents — A.8.7's allowlisting (a) and malware detection (f) can prevent standalone port-knocking malware from running, but the technique can also be implemented via legitimate host-based firewall rules, raw-socket abuse by already-permitted processes, or built into authorized software — none of which A.8.7 reaches.
- T1205.001partialresponds — Anti-malware software can quarantine and eradicate the custom malicious software or malware used to implement the port-knocking listener, though it cannot remediate port knocking implemented through native host-based firewall rules.
- T1205.002partialdetects — A.8.7 mandates malware detection and repair software that scans data received over networks and files on electronic storage media, which can surface the crafted packet or the backdoor it activates, but the technique's low CPU overhead and limited visibility into raw socket usage leave a nameable remainder outside the control's required scanning scope.
- T1205.002partialresponds — A.8.7's malware repair capabilities can quarantine or remove the malicious implants that install socket filters, though it may not remediate filters established by abusing legitimate system tools.
- T1210partialprevents — A.8.7 reduces the attack surface through vulnerability management (c) and application allowlisting (a), but exploitation of a zero-day or unpatched remote service remains possible, so the technique is constrained rather than eliminated.
- T1211partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, webpages and storage media, which surfaces exploitation artifacts delivered as malware, but the technique's core evasion—exploiting vulnerabilities in logging pipelines or security tools to suppress telemetry—occurs outside the malware scanner's view, leaving a bounded remainder.
- T1211partialprevents — A.8.7's malware detection and repair software, application allowlisting, and vulnerability reduction controls can prevent the initial delivery or execution of malware that would be used to exploit a vulnerability for stealth, but the technique can also be performed by an already-present actor exploiting a vulnerability directly without malware delivery, which these controls do not address.
- T1212partialprevents — A.8.7(c) reduces vulnerabilities exploitable by malware and (f) detects malicious payloads, which catches a slice of T1212 where the exploit chain involves malware or malware-exploitable flaws, but the control is framed for malware rather than credential-access exploitation generally, so exploits that bypass malware delivery (e.g., direct memory reads, token forgery via Kerberos flaws) are not addressed.
- T1216partialdetects — A.8.7's malware detection software, allowlisting and scanning can surface some instances of T1216 (e.g. when the proxied payload is a scannable file, or when behavioral/heuristic layers in the detection stack flag a trusted script spawning an unexpected child), but the technique is explicitly designed to bypass signature validation and application control by abusing Microsoft-signed, default-on-Windows scripts, so the primary mechanisms A.8.7 mandates miss the core proxy-execution pattern.
- T1216partialprevents — A.8.7's allowlisting (a), malware detection (f), and vulnerability reduction (c) can constrain T1216, but the technique is explicitly designed to bypass application control and signature validation by abusing trusted, signed scripts — the very mechanisms A.8.7 prescribes — so what is prevented is a slice rather than the bulk.
- T1216partialresponds — A.8.7's malware repair software can quarantine the malicious payload and terminate the execution, but the legitimate system script abused to proxy the execution is not removed.
- T1216.001partialdetects — A.8.7's malware detection software (f) would scan the downloaded .sct file or its executed payload, and blocklisting (b) could catch the remote hosting URL, but the technique is explicitly designed to bypass signature validation and application control solutions that do not account for PubPrn abuse, leaving a named gap in what the control's detection mechanisms reach.
- T1216.001partialprevents — A.8.7's allowlisting (8.7a), malicious-website blocklisting (8.7b), and malware detection (8.7f) can constrain PubPrn abuse, but the technique is explicitly designed to bypass application control and signature validation by abusing a Microsoft-signed system script, and PubPrn.vbs is a legitimate component that cannot be removed without breaking printing — so what is prevented is a slice, not the bulk.
- T1216.001partialresponds — Malware repair software can contain the attack by terminating the malicious process and eradicating the downloaded payload, though it does not remove the legitimate PubPrn script itself.
- T1216.002partialdetects — A.8.7's malware detection software (item f) relies primarily on signature-based scanning, which will not flag a legitimate, Microsoft-signed script like SyncAppvPublishingServer.vbs; behavioral or EDR-style monitoring of the wscript.exe → SyncAppvPublishingServer.vbs → powershell.exe chain can surface the abuse, but A.8.7 does not mandate that depth of instrumentation, and the technique is explicitly designed to evade defensive countermeasures.
- T1216.002partialprevents — A.8.7's allowlisting (a) and malware detection (f) provide some protection — blocking wscript.exe or detecting the malicious PowerShell payload — but SyncAppvPublishingServer.vbs is a legitimate, Microsoft-signed System32 binary that allowlisting would typically permit, and the technique is specifically designed to bypass these controls by proxying execution through a trusted LOLBin.
- T1216.002partialresponds — Malware repair software can terminate the malicious PowerShell process or quarantine the payload executed via the proxy, but it cannot eradicate the legitimate SyncAppvPublishingServer script itself.
- T1218partialdetects — A.8.7's malware detection software (clause f) is largely signature-based, and T1218 is explicitly designed to bypass signature-based defenses by proxying through trusted, Microsoft-signed binaries; however, behavioral detection, EDR-style monitoring, allowlisting deviations (clause a), and scanning of proxied content can surface some instances, so detection is real but bounded by the technique's evasion design.
- T1218partialprevents — A.8.7's allowlisting (a), blocklisting (b), defense-in-depth scanning (f, g), and vulnerability reduction (c) can prevent some instances of T1218, but the technique is specifically designed to bypass signature-based and process-based defenses by proxying through trusted, often Microsoft-signed binaries that allowlisting typically permits and that signature scanning explicitly does not flag.
- T1218partialresponds — Malware repair software can terminate the malicious process instance and quarantine the payload being executed, but cannot eradicate the trusted system binary being abused as a proxy.
- T1218.001partialprevents — A.8.7 requires application allowlisting and controls that prevent unauthorized software, which can block hh.exe or .chm execution on hardened systems, but the technique also relies on user execution and legacy bypasses that the control's awareness and configuration guidance cannot fully close.
- T1218.001partialresponds — Malware repair software can quarantine malicious CHM files and kill the associated processes, eradicating the artifact when the payload is recognized as malware.
- T1218.002partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can surface a malicious .cpl file at rest or in transit, but the technique's abuse of a signed Microsoft binary for proxy execution may evade behavioural detection, leaving a gap.
- T1218.002partialprevents — A.8.7's allowlisting (a), malware detection (f), defense-in-depth (g), and email/attachment scanning (f.2) can prevent some instances by blocking unauthorized CPL files or detecting malicious payloads before they reach control.exe, but the technique is explicitly designed to bypass application/file-extension allowlists and abuses a legitimate Windows system binary that cannot be removed without breaking the platform, leaving a bounded remainder where the proxy execution still succeeds.
- T1218.003partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, and CMSTP.exe loading a malicious DLL or scriptlet from a remote server creates artifacts (the INF file, the network connection, the loaded module) that endpoint or gateway scanning can surface, but the technique's use of a signed Microsoft binary and potential for encrypted channels limits detection to implementations with sufficient instrumentation depth, leaving a nameable remainder.
- T1218.003partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block CMSTP.exe or scan malicious INF files, but CMSTP is a Microsoft-signed binary that the technique explicitly bypasses AppLocker with, and the binary may be needed for legitimate Connection Manager profiles, so what is prevented is a slice rather than the bulk.
- T1218.004partialdetects — A.8.7 mandates malware detection and repair software that scans files and data content, which can surface a malicious .NET binary when it is written to disk or loaded, but the control's scanning scope does not guarantee runtime observation of a signed Microsoft utility executing a malicious installer component in memory.
- T1218.004partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block InstallUtil-based proxy execution where configured to do so or where the malicious .NET payload is detected, but InstallUtil is a legitimate signed Microsoft binary that cannot be universally removed without breaking .NET operations, and the technique is explicitly designed to bypass application control via trusted-binary proxying — so what is prevented is a slice, not the bulk.
- T1218.004partialresponds — Malware repair software can terminate the malicious process and quarantine the associated .NET payload, though it cannot remove the trusted InstallUtil utility itself.
- T1218.005partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block or catch some mshta.exe abuse, but mshta.exe is a legitimate Windows utility that often cannot be removed without breaking functionality, and the technique is explicitly designed to bypass application control that does not account for its abuse — so what is prevented is a slice, not the bulk.
- T1218.005partialresponds — A.8.7's malware detection and repair software (clause f) can detect and quarantine the malicious payload delivered via mshta.exe, and clause (k) supports isolation of environments — both are genuine containment/eradication actions once the technique is underway. However, A.8.7 is not a dedicated incident response control: it does not assess whether an event is an incident, collect evidence, or coordinate the broader response process, and its eradication is limited to what the AV/EDR automatically removes rather than the full scope of response tasks.
- T1218.007partialprevents — A.8.7's allowlisting (8.7a) and malware scanning of files (8.7f) can prevent some msiexec abuse, but msiexec.exe is a required, signed Windows binary that the technique is specifically designed to abuse, so it cannot be removed and basic allowlists miss it — the same shape as cm-7 vs T1059.
- T1218.008partialdetects — A.8.7's malware detection software can surface the malicious DLL payload or the suspicious odbcconf.exe command line, but the technique is explicitly designed to bypass application control by abusing a signed Microsoft binary, so detection rests on catching the payload or behavior rather than the technique itself, leaving a meaningful slice undetected.
- T1218.008partialprevents — A.8.7's application allowlisting guidance (a) and malware detection (f) can block odbcconf.exe abuse, but the binary is a signed Microsoft utility that may be required for legitimate operations and the technique is specifically designed to bypass application control that does not account for its abuse, leaving a bounded remainder the control cannot reach.
- T1218.009partialprevents — A.8.7's allowlisting guidance (a) and malware detection (f) can prevent execution where Regsvcs/Regasm are not on the allowlist or where the malicious .NET assembly is detected, but these are legitimate Microsoft-signed utilities often required by the platform and applications, and the technique is specifically designed to bypass application control via signed-binary proxy execution, so what is prevented is a slice rather than the bulk.
- T1218.010partialprevents — A.8.7's allowlisting (8.7a) cannot block regsvr32.exe itself because it is a Microsoft-signed binary required for normal Windows operations, and the technique is explicitly designed to evade allowlists; URL blocklisting (8.7b) and malware detection (8.7f) can catch some payloads and remote sources, but the technique is built to bypass these via COM scriptlets and signed-binary proxying, leaving a bounded remainder that the control cannot reach.
- T1218.010partialresponds — A.8.7's malware detection and repair software (item f) can quarantine the malicious DLL/scriptlet and terminate the Regsvr32 process once the payload is observed executing, which is containment and eradication of the technique underway; however, the clause's response capability is bounded to what endpoint protection can automate and does not extend to the broader incident response activities (assessment, evidence collection, root cause analysis) that A.5.26 covers, leaving a named remainder.
- T1218.011partialdetects — A.8.7's malware detection and repair software scans files, attachments, downloads and webpages at multiple points (gateways, endpoints, servers), so malicious DLLs and scripts loaded via rundll32 are subject to scanning; however, the technique is explicitly designed to evade security tools by abusing a trusted Windows binary, and variants using legitimate signed system DLLs (zipfldr.dll, ieframe.dll) carry no malicious payload for signature-based detection to catch, leaving behavioral and content-based detection as a slice rather than a complete mechanism.
- T1218.011partialprevents — A.8.7's allowlisting (a) and malware detection (f) can prevent unauthorized DLLs from being loaded by rundll32.exe, but the technique is explicitly designed to evade these controls by abusing a legitimate, signed system binary that allowlists typically permit, so prevention is a slice rather than the bulk.
- T1218.011partialresponds — A.8.7(f) malware detection and repair software detects and removes malicious payloads that have already entered the system, and (k) isolates environments where catastrophic consequences can occur — both are response actions against a technique that has run. However, A.8.7 is primarily preventive, and T1218.011 specifically abuses the signed rundll32.exe binary to evade detection (acknowledged in the clause's own guidance at (g)(2)), so the response reach is bounded to what the detection software can catch after the fact, not a comprehensive containment and eradication capability.
- T1218.012partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can surface verclsid.exe proxying a malicious COM payload when the payload is written to disk or traverses a monitored channel, but the control's scope does not guarantee runtime behavioral detection of a signed, native binary executing a scriptlet from a remote server.
- T1218.012partialprevents — A.8.7's allowlisting (a) and malware detection (f) provide some protection, but verclsid.exe is a signed native Windows binary that allowlists typically permit — the technique is specifically designed to bypass application control via this proxy — so the proxy execution itself is not prevented, though the malicious COM payload it loads may be detected by malware scanning.
- T1218.013partialdetects — A.8.7's malware detection software is the very class of 'security products' the technique is documented to evade (signed Microsoft binary proxying execution), so signature-based AV largely misses it; however, behavioral EDR and allowlisting (item a) can surface the abuse, leaving detection as a slice rather than a bounded remainder of a working mechanism.
- T1218.013partialprevents — A.8.7's allowlisting guidance (8.7a) can block mavinject.exe where the organization does not need App-V functionality, but mavinject.exe is a legitimate, Microsoft-signed Windows utility that many environments require, and the technique is explicitly designed to evade detection by proxying through a signed binary — so what is prevented is a slice, not the bulk, mirroring the cm-7 vs T1059 shape where required platform utilities cannot be removed and adversaries preferentially abuse them.
- T1218.013partialresponds — Malware repair software can terminate the injected process and quarantine the malicious DLL, providing partial response by eradicating the payload while leaving the legitimate mavinject.exe utility intact.
- T1218.014partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can detect malicious .msc files on disk or in transit, but the control's scope is malware-specific and does not guarantee detection of all signed binary proxy execution techniques.
- T1218.014partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block or scan malicious .msc files and constrain some delivery paths, but mmc.exe is a required, signed Microsoft binary that cannot be removed from the build — analogous to the cm-7 vs T1059 anchor — and the technique's abuse of legitimate snap-ins and CLSIDs evades the bulk of what the clause's mechanisms can stop, leaving a bounded remainder that runs.
- T1218.014partialresponds — A.8.7's malware detection and repair software (clause f) and business continuity/recovery procedures (clauses j, l) act on malware once it has entered the system, which is a form of containment and eradication; however, the clause's primary function is prevention and detection, and its response to T1218.014 is indirect — it addresses the resulting malware payload rather than the MMC abuse technique itself, and does not constitute the full incident-response procedure that A.5.26 provides.
- T1218.015partialdetects — A.8.7's detection software and scanning can surface anomalous child-process behaviour from legitimate Electron apps (e.g., chrome.exe spawning cmd.exe) where behavioural or EDR-style telemetry is in scope, but the technique is specifically designed to evade signature-based AV by abusing trusted signed binaries, and application allowlisting does not help because the host applications are legitimate — so what is caught is a slice dependent on implementation depth rather than the bulk of the technique.
- T1218.015partialprevents — A.8.7's allowlisting (8a) cannot prevent this because the technique abuses legitimate, authorized Electron applications (Teams, Slack, Chrome); however, malware detection software (8f) and defense-in-depth (8g) can partially prevent by scanning for malicious JavaScript planted in Electron apps, detecting suspicious child process command lines (e.g., --gpu-launcher pointing to cmd.exe), and blocking known malicious patterns, though the core abuse of trusted signed binaries largely bypasses traditional prevention.
- T1218.015partialresponds — Malware repair software can quarantine malicious JavaScript planted within Electron applications or payloads executed by them, eradicating the file-based artifacts of the technique, though it may not address command-line abuse of the legitimate binaries themselves.
- T1219partialdetects — A.8.7 detects some instances of T1219 via application allowlisting (8.7a) catching unauthorized RAT installations and via behavioral detection in modern EDR/malware tools, but the technique's core is abusing LEGITIMATE remote access tools (TeamViewer, AnyDesk, EDR remote features) that malware detection software is not designed to flag — the clause's primary mechanism (malware detection and repair software, 8.7f) targets malicious code, not abuse of authorized software, leaving the bulk of the technique undetected.
- T1219partialprevents — Application allowlisting (clause a) and blocklisting (b) can prevent unauthorized remote access tools from running, but T1219 specifically abuses legitimate remote access software (TeamViewer, RDP, Chrome Remote Desktop, EDR response features) that the organization typically needs and cannot disallow — the technique's core is the abuse of authorized tools, which allowlisting cannot reach.
- T1219.002partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, but remote desktop software is legitimate signed software that malware scanners do not flag, so detection depends on the behavioural monitoring and network gateway scanning the control also requires, which is a slice of the control's mechanisms rather than its primary one.
- T1219.002partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can block unauthorized remote desktop tools, but the technique explicitly notes these tools 'may be allowed by application control' as legitimate support software, and malware detection will not flag legitimate binaries — so the abuse of authorized RMM tools, which is the core of T1219.002, is not prevented.
- T1220partialdetects — A.8.7's malware detection software (item f) can scan XSL files for embedded malicious scripts and EDR components may flag msxsl.exe or wmic /FORMAT spawning behavior, but the technique is specifically designed to evade detection by leveraging trusted signed Microsoft binaries and valid XML/XSL files that often evade signature-based scanning, so detection is a slice rather than a bounded remainder of a working mechanism.
- T1220partialprevents — A.8.7's allowlisting (a) and malware detection (f) can block the msxsl.exe variant and detect malicious XSL content, but the technique is explicitly designed to bypass application control, and the Squiblytwo variant abuses wmic.exe — a built-in Windows utility that cannot be allowlisted out without breaking system management — so prevention is partial and not mostly.
- T1220partialresponds — Malware protection software can quarantine the dropped msxsl.exe utility or the malicious XSL files once detected, eradicating the specific artifacts used in the execution.
- T1221partialdetects — A.8.7 explicitly mandates scanning email attachments (f.2), scanning webpages when accessed (f.3), and network gateway detection (g.1), all of which are relevant detection points for template injection; however, the technique is specifically designed to evade static detection since the malicious payload is fetched remotely and no typical indicators are present in the document until after fetch, so detection depends on whether the AV/EDR recognizes the template reference pattern or the remote payload, leaving a bounded remainder.
- T1221partialprevents — A.8.7 has multiple mechanisms that can partially prevent template injection — blocklisting the remote URL the template references (b), scanning the fetched payload as a webpage (f)(3), application allowlisting (a), and user awareness (m) — but the technique is explicitly designed to evade static detection since the malicious payload is fetched at runtime from a URL, leaving a bounded remainder where new/unknown URLs, encrypted payloads, and legitimate-looking infrastructure slip past.
- T1221partialresponds — Malware repair software can quarantine the malicious document or the downloaded payload to contain the threat, but this does not address template injections used to force authentication attempts.
- T1222partialdetects — A.8.7's malware detection software can catch known malware that performs T1222 via signatures, and allowlisting (item a) plus integrity validation (item d) can surface some unauthorized changes, but T1222 is typically executed through legitimate system utilities (icacls, chmod, chown) that AV does not flag, so the technique itself is largely invisible to the control's primary mechanism.
- T1480.001partialdetects — A.8.7's malware detection software is precisely what environmental keying is designed to evade — the technique explicitly aims to defeat 'anti-virus detection' by encrypting the payload so it only decrypts in target environments, leaving sandboxes and AV inert. However, A.8.7's broader scope (behavioral monitoring, the specific guidance on evasive techniques in (g)(2), detection of environmental probing or post-execution behavior in the target) catches a slice of the technique, so the verb holds but only partially.
- T1480.001partialprevents — A.8.7's allowlisting (a) and defense-in-depth (g) can prevent the decrypted payload from executing, but environmental keying is specifically designed to evade the clause's primary mechanism — malware detection software (f) sees an encrypted payload that does not match signatures — so the core evasion technique itself is not prevented, only the downstream execution is constrained where allowlisting is in force.
- T1480.002partialprevents — A.8.7's allowlisting, scanning and detection mechanisms can prevent the malware from executing in the first place, which prevents the mutex from ever being created or checked; but the mutex itself is a legitimate OS synchronization primitive that A.8.7 does not specifically target, so prevention is indirect and depends on stopping the broader malware execution rather than the mutex behavior itself.
- T1480.002partialdetects — A.8.7's malware detection software (item f) and its explicit awareness of evasive techniques (item g.2) provide some detection of mutex-based behaviors via behavioral analysis and known mutex-name signatures for established malware families, but mutexes are legitimate OS constructs and the technique is designed to be stealthy with dynamically generated names, so detection covers a slice rather than the bulk.
- T1485partialdetects — A.8.7's malware detection and repair software, application allowlisting, and behavioral scanning would surface the malware performing T1485 and the mass file-overwrite activity itself in most cases, though sophisticated wiper malware (Shamoon-class) using novel or encrypted payloads can evade signature-based detection, leaving a bounded remainder.
- T1485partialprevents — A.8.7 prevents the subset of T1485 delivered via malware that detection/allowlisting catches, but the technique also executes through legitimate OS commands (del, rm), cloud API calls, and insider use of valid credentials — none of which A.8.7 addresses — and the clause itself states that anti-malware alone is not usually adequate.
- T1485partialresponds — Malware protection and repair software can quarantine wiper malware and kill its processes to contain the destruction, but this provides no response against data destruction performed via native operating system commands or cloud API calls.
- T1486partialresponds — A.8.7's malware detection and repair software (clause f) can detect ransomware encryption behavior in flight and quarantine or terminate the malicious process, which is a containment/eradication action; clause (k) on isolating environments also supports containment. However, the control does not perform assessment of whether an event is an incident, evidence collection, forensic investigation, or removal of the actor's foothold — the encryption already realized is not undone, and lateral movement/credential-based propagation are outside the malware-scanner's scope.
- T1490partialdetects — A.8.7's malware detection and repair software, particularly modern EDR with behavioral monitoring, can flag T1490 actions (e.g., vssadmin delete shadows, wbadmin delete catalog, bcdedit recovery disable) as suspicious behaviors, and item (d)'s automated validation of system content can surface unauthorized amendments to recovery features; however, the control's primary focus is detecting malware itself rather than detecting abuse of legitimate system utilities, and many conformant implementations (signature-based AV, allowlisting focused on unauthorized software) would not catch these techniques.
- T1490partialresponds — A.8.7 explicitly requires business continuity plans for recovering from malware attacks (including backups and recovery measures) plus procedures for reporting and recovering from malware attacks; T1490 is the exact technique used by ransomware (a malware attack) to block those recovery options, so the control's response actions engage once the malware event is underway.
- T1490partialprevents — A.8.7's anti-malware controls (detection software, allowlisting, blocklisting, vulnerability management) can prevent the malware that would execute T1490 from running in the first place, but the technique uses built-in Windows utilities (vssadmin, wmic, wbadmin, bcdedit) that are typically permitted, and A.8.7 does not specifically protect recovery features from being disabled — so prevention is incidental rather than targeted, and partial rather than mostly.
- T1491partialdetects — A.8.7's malware detection software and clause (d)'s detection of unapproved files or unauthorized amendments can surface defacement when it is delivered via malware or involves unauthorized file changes, but defacement frequently occurs through non-malware vectors (direct access, SQL injection, web shells) that fall outside A.8.7's malware-centric scope.
- T1491.001partialdetects — A.8.7's guidance (d) on regular automated validation of software and data content and investigation of unapproved files or unauthorized amendments directly addresses the file/content modifications that internal defacement performs, and malware detection (f) can catch tools used to enable it; however, the control's primary scope is malware rather than content integrity, and defacement performed through legitimate system tools after access is gained falls outside its detection mechanism.
- T1491.001partialprevents — A.8.7's allowlisting (a), vulnerability reduction (c), and malware detection (f) can block some paths an adversary uses to reach and modify internal systems, but defacement is a post-compromise modification often performed with legitimate system tools (PowerShell, RDP, file edits) that are not malware, so the control's mechanism does not target the technique itself and leaves substantial attack surface intact.
- T1495partialprevents — A.8.7's malware detection software, vulnerability management (8.7c), and application allowlisting (8.7a) can prevent some firmware-corruption vectors — catching the malware that performs the flashing, patching exploitable firmware flaws, and blocking unauthorized flashing tools — but the clause's core mechanism is file/email/web scanning aimed at software-level malware, not firmware integrity verification, leaving substantial firmware-level attack paths unaddressed.
- T1496partialdetects — A.8.7's malware detection software and allowlisting catch a meaningful slice of resource hijacking — known cryptominers, proxyjacking implants, and similar malware payloads are within its detection scope — but resource hijacking frequently uses legitimate binaries, fileless techniques, or cloud-API abuse that produces no malware artifact for the control to scan, leaving that portion undetected.
- T1496partialprevents — A.8.7's malware detection and repair software, application allowlisting, and malicious website blocking can prevent cryptominers and proxyjacking malware from executing, but resource hijacking via compromised SaaS credentials or abuse of cloud messaging services bypasses endpoint malware controls entirely.
- T1496partialresponds — Malware repair software can quarantine and remove cryptominers or other malicious payloads executing on endpoints, but this eradicates only the subset of resource hijacking that relies on identifiable malware, leaving SaaS abuse and non-malware hijacking unaddressed.
- T1496.001partialdetects — A.8.7's malware detection and repair software (f), application allowlisting (a), and regular automated validation of software content (d) all run on the organization's own systems where compute hijacking executes, and cryptocurrency miners are a well-known malware class that endpoint protection routinely detects; the bounded remainder is fileless miners, LOLBin-based miners, and containerized deployments that evade signature-based detection.
- T1496.001partialprevents — A.8.7's allowlisting and vulnerability reduction can block the initial malware delivery or exploitation that leads to compute hijacking, but once the malware is executing, the control's preventive measures are bypassed and the resource consumption continues.
- T1496.001partialresponds — A.8.7's malware detection and repair software (clause f) does detect and remove crypto-mining malware once it is running on a system, which is a genuine containment/eradication action against a technique that has executed; however, A.8.7 is primarily a protection control whose response capability is one component among many preventive measures, and the response is bounded by what the detection software can identify (e.g., evasive or fileless miners may evade it), so it does not reach the `mostly` rung that the dedicated incident-response control A.5.26 reaches on comparable malware techniques.
- T1496.002partialdetects — A.8.7's malware detection and repair software (8.7f) and application allowlisting (8.7a) can surface the malware that performs bandwidth hijacking when it is a known or unauthorized binary, but the technique explicitly includes proxyjacking via legitimate proxyware services, living-off-the-land tools, and evasive malware that bypass signature-based detection, so a substantial share of instances produce no malware artifact for A.8.7 to detect.
- T1496.002partialprevents — A.8.7 prevents T1496.002 only transitively, by blocking the malware that would hijack bandwidth; the technique is specifically designed to be stealthy (botnets, proxyjacking) and A.8.7's own guidance acknowledges evasive techniques, so prevention is a slice rather than a closure.
- T1496.002partialresponds — A.8.7's malware repair software and isolation procedures contain and eradicate the malicious payloads driving the bandwidth hijacking, leaving a remainder of non-malware tools or legitimate proxyware abused by the adversary.
- T1497partialdetects — A.8.7's defense-in-depth, behavioral analysis, and regular automated validation (clauses d, g) can surface some indicators of sandbox/VM checks, but T1497 is specifically designed to evade the very analysis environment A.8.7's detection and repair software creates — the malware disengages or hides its true functionality when VM artifacts are detected, so the core scanning mechanisms (clause f) largely miss the evasion behavior itself.
- T1497partialresponds — A.8.7 includes malware repair software that quarantines and removes malicious payloads, which eradicates the malware performing the sandbox evasion once it is identified.
- T1497.001partialprevents — A.8.7's allowlisting and malware detection can prevent the implant from executing at all, which prevents its system checks from running; however, T1497.001 is specifically an evasion technique designed to defeat analysis environments, and sophisticated malware that bypasses A.8.7's detection will still perform these checks unimpeded, leaving a substantial remainder.
- T1497.001partialdetects — A.8.7's malware detection software can detect the malware payload that performs system checks (signature-based detection), but the system check technique itself involves legitimate system queries (WMI, PowerShell, registry, hardware checks) designed to look like normal activity, and A.8.7 does not specifically target these evasion behaviors — detection is indirect and depends on catching the payload rather than the check behavior.
- T1497.001partialresponds — Malware repair software can quarantine or terminate the malicious process performing the system checks, but because these checks often rely on native OS utilities (such as WMI or PowerShell), automated eradication may not trigger unless a known malware signature is matched.
- T1497.003partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, and time-based evasion checks are a known malware technique that updated signatures and heuristics can flag, but the control's scanning scope is file and content inspection rather than runtime behavioral analysis of API calls like GetTickCount, so detection is a slice the implementer's tooling may or may not cover.
- T1505partialdetects — A.8.7's item (d) explicitly calls for automated validation investigating 'unapproved files or unauthorized amendments' on critical systems, and item (f) malware detection can surface known malicious server components, so detection mechanisms exist for T1505; however, the technique abuses legitimate extensible features with custom or legitimate-looking payloads that evade signature-based scanning, and the control's primary focus is broader malware rather than server-component abuse specifically, leaving a substantial slice undetected.
- T1505partialprevents — A.8.7's allowlisting (a), integrity validation of software and data content (d), and malware detection (f) can prevent some instances of malicious server components, but T1505 specifically abuses legitimate extensible features (web shells, SQL procedures, plugins) that are often required for the application to function and may not be flagged by allowlists or signature-based detection, leaving a substantial slice reachable.
- T1505partialresponds — Malware repair software can quarantine or delete known malicious server components (such as web shells or malicious plugins), but it may not revert the configuration changes made to the server application to load them.
- T1505.002partialdetects — A.8.7's malware detection software, application allowlisting (item a), and automated validation of software content (item d) could surface a malicious .NET transport agent assembly, but the technique specifically abuses a legitimate Exchange feature and is designed to be stealthy within trusted infrastructure, so detection is not guaranteed and depends on implementation choices rather than the clause's mechanism.
- T1505.002partialprevents — A.8.7's application allowlisting (a), software validation (d), and malware detection (f) can constrain unauthorized .NET assemblies from being registered as Exchange transport agents, but the technique abuses legitimate Exchange functionality, requires Exchange admin access (outside A.8.7's scope), and the clause is general malware protection rather than Exchange-specific hardening, so prevention is a slice rather than comprehensive.
- T1505.002partialresponds — Malware repair software can quarantine the malicious .NET assembly file used as the transport agent, though it may not revert the agent's registration within Exchange.
- T1505.003partialprevents — A.8.7 contributes to prevention through application allowlisting (a), vulnerability management (c), and protective measures for files from external networks (e), but web shells use legitimate scripting languages that web servers require, are often placed via web application exploitation not directly addressed by this clause, and evade traditional malware detection, so prevention is partial rather than complete.
- T1505.003partialresponds — A.8.7's malware detection and repair software (f), integrity validation of software and data content (d), and isolation of environments (k) provide genuine response actions against a web shell already placed on a server — detection, removal, and containment of the compromised host. But A.8.7 is primarily a preventive control; it lacks the full incident response procedure (assessment, evidence collection, root cause analysis, eradication strategy) that A.5.26 carries, so the response share is a slice rather than the bulk.
- T1505.004partialdetects — A.8.7's allowlisting (a), regular automated validation of software content (d), and malware detection/repair software (f) can surface unauthorized DLLs and unapproved files on IIS servers, but the clause is general-purpose anti-malware rather than IIS-specific: signed or legitimately-named modules, custom payloads without signatures, and components that evade file-based scanning can persist undetected, and behavioral detection of in-process IIS module abuse is not mandated.
- T1505.004partialprevents — A.8.7's guidance (a) on application allowlisting and (d) on regular automated validation of software content directly addresses unauthorized DLL installation, and (f) malware detection can scan for malicious IIS components, but the technique presupposes existing access for persistence, IIS must load legitimate DLLs so allowlisting cannot be fully closed, and sophisticated components evade detection — so prevention is real but bounded.
- T1505.004partialresponds — A.8.7's malware detection and repair software (clause f) and its 'repair' function explicitly perform eradication — detecting and removing malicious IIS components once installed — and clause (k) supports isolation of compromised environments, so a response action exists; however, the clause is primarily preventive, the repair software is not specifically aimed at IIS persistence mechanisms, and sophisticated ISAPI/IIS modules can evade signature-based detection, leaving the response as a slice rather than comprehensive containment and eradication.
- T1505.005partialdetects — A.8.7(d) explicitly addresses investigating 'unapproved files or unauthorized amendments' to software, which is precisely what T1505.005 performs when modifying termsrv.dll or its ServiceDll registry value, so file integrity monitoring implementations can surface this. However, the technique abuses a legitimate Windows system DLL and is designed to be subtle, and the control's primary mechanism — malware detection and repair software per (f) — frequently does not flag modifications to legitimate system components, so detection is a slice rather than the bulk.
- T1505.005partialprevents — A.8.7's file-integrity validation (d), malware detection (f), and application allowlisting (a) can catch some instances — a separately dropped malicious DLL is scannable, and unauthorized amendments to termsrv.dll or its ServiceDll value are investigable under (d) — but the technique specifically abuses a legitimate Windows component, so allowlisting does not block it and standard malware detection is not reliably triggered by a patched termsrv.dll or by a wrapper DLL that mimics expected behaviour, leaving a meaningful remainder.
- T1505.006partialdetects — A.8.7's broader mechanisms — particularly (d) regular automated validation of software content and (a) application allowlisting — could surface unauthorized VIBs on ESXi if those systems are in scope, but the clause's primary mechanism (f) malware detection and repair software does not run on ESXi hypervisors, and VIBs are signed legitimate packages rather than traditional malware, so signature-based detection would not catch them.
- T1505.006partialprevents — A.8.7's general malware protection guidance (allowlisting of unauthorized software, malware detection, defense in depth) could partially address malicious VIB installation, but the technique operates at the ESXi hypervisor level which is typically outside the scope of traditional endpoint malware protection, and VIBs are a legitimate VMware mechanism that allowlisting may not block without specific ESXi-level configuration.
- T1505.006partialresponds — A.8.7 includes malware repair and environment isolation which provide eradication and containment, but standard anti-malware tools often lack the capability to automatically remove malicious vSphere Installation Bundles from ESXi hypervisors, leaving a significant manual remediation remainder.
- T1525partialdetects — A.8.7's malware detection and repair software (f) and software-content validation (d) can scan images and detect unauthorized amendments, and allowlisting (a) can flag unapproved images, but the technique targets cloud/container images specifically — a domain where traditional endpoint malware scanning often has limited reach, and subtle implants (web shells, backdoors) may evade signature-based detection.
- T1525partialresponds — Anti-malware repair functions can quarantine or delete infected files to eradicate the malicious code, but fully remediating a compromised container registry typically requires registry-specific image deletion and pipeline remediation beyond standard malware repair.
- T1534partialdetects — A.8.7's malware detection software scans email attachments, downloads and links (clause f), which surfaces malware-bearing payloads delivered via internal spearphishing, but the technique's credential-harvesting links, chat-app lures (Teams etc.), and social-engineering-only variants produce no malware artifact for the scanner to flag, and internal-originated mail often bypasses gateway scrutiny.
- T1534partialprevents — A.8.7's email/attachment scanning (f.2), web scanning (f.3), and application allowlisting (a) can catch malicious payloads and links delivered via internal spearphishing, but the technique is specifically designed to leverage already-trusted internal accounts, can succeed through pure credential harvesting on fake login pages without malware, and increasingly uses chat platforms (Teams, Slack) that fall outside the email-scanning scope; the trust exploitation that defines T1534 is not addressed by malware protection controls.
- T1534partialresponds — A.8.7's malware detection and repair software (clause f) can detect and clean malicious payloads delivered via internal spearphishing, and its isolation guidance (clause k) can contain environments where malware has spread — both genuine response actions once the technique is underway. However, A.8.7 does not address the compromised account or the spearphishing campaign itself (containing the actor's foothold, revoking the abused credentials), which fall under A.5.26 incident response and A.8.5 authentication. The response is therefore limited to the malware-delivery aspect of the technique, not the full campaign.
- T1539partialdetects — A.8.7's malware detection software and webpage/attachment scanning catch the variants of T1539 that rely on known malware (on-disk cookie stealers, malicious attachments), but the technique is also executed via JavaScript injection, AitM proxies (Evilginx2/Muraena), and user-executed scripts that are not 'malware' in the sense A.8.7's mechanism targets — so detection covers a slice, not the bulk.
- T1539partialprevents — A.8.7 partially prevents T1539: malware detection and repair software (clause f) can catch some cookie-stealing malware, blocklisting (b) and webpage scanning (f.3) address malicious-website vectors, and awareness training (m) helps against user-execution paths — but the technique's adversary-in-the-middle vectors (Evilginx2, Muraena) involve no malware on the victim system, and sophisticated cookie stealers frequently evade signature-based detection, leaving a substantial portion of the technique's surface unaddressed.
- T1539partialresponds — A.8.7 provides containment and eradication (repair and isolation) for malware, which bounds and removes the threat when the cookie theft is performed by an infostealer, but leaves non-malware vectors like malicious proxies and injected JavaScript unaddressed.
- T1542.001partialdetects — A.8.7 mandates malware detection and repair software with regular scanning of files, email, and web content, which can detect malicious firmware modification tools or payloads at rest or in transit, but the control's guidance focuses on software and data content scanning rather than firmware integrity verification, leaving firmware-resident malware outside its primary detection scope.
- T1542.003partialdetects — A.8.7's malware detection software and the (d) validation clause can surface bootkit artifacts (MBR/VBR/ESP modifications, post-OS payload behaviour) on systems where such checks are implemented, but the clause's primary mechanisms — file, email and web scanning, allowlisting, blocklisting — operate at and above the OS layer, while a bootkit by definition executes below it; the boot sector layer is not what the clause is built to see, so what is caught is a slice rather than the technique.
- T1542.003partialprevents — A.8.7's allowlisting (a), vulnerability reduction (c), and defense-in-depth placement (g) can constrain some bootkit delivery vectors, but the technique executes below the OS layer where the clause's detection/repair software (f) and most other controls operate, so a bootkit that has already written to MBR/VBR/ESP runs before A.8.7's mechanisms engage.
- T1542.003partialresponds — A.8.7 provides response capabilities for malware incidents — isolation of environments (k), business continuity/recovery plans (j), and procedures for dealing with malware attacks including reporting and recovering (l) — which apply once a bootkit has established persistence. However, bootkits operate below the OS layer (MBR/VBR/ESP), so the clause's primary mechanism — malware detection and repair software (f) — is largely ineffective at detecting or eradicating them, and full remediation typically requires re-imaging or specialized tools outside A.8.7's scope.
- T1543partialdetects — A.8.7's malware detection software and the clause (d) automated validation that investigates unapproved files and unauthorized amendments can surface malicious service installations and known-malware payloads, but T1543 is frequently executed with built-in system tools (sc.exe, systemctl, launchctl) carrying no traditional malware signature, so a meaningful slice of the technique runs without triggering the control's detection mechanisms.
- T1543partialprevents — A.8.7's allowlisting (a) and automated validation of software/content investigating unapproved files and unauthorized amendments (d) directly constrain creation of new unauthorized services, daemons and agents, but T1543 also covers MODIFYING existing services that are already allowlisted, and privileged actors can bypass these controls, leaving a meaningful remainder.
- T1543partialresponds — Malware repair software can quarantine malicious executables and remove associated service registrations, eradicating the persistence mechanism, but this does not cover services modified to run legitimate administrative tools or scripts.
- T1543.001partialdetects — A.8.7 mandates regular automated scanning of systems and files, which can surface unauthorized .plist files in LaunchAgent directories, but the control's scope is determined by risk assessment and does not guarantee host-level file integrity monitoring depth.
- T1543.001partialprevents — A.8.7's allowlisting (item a) and detection/validation (items d, f) can prevent or catch some unauthorized Launch Agent installations, but Launch Agents are a legitimate macOS mechanism used pervasively by benign software, so they cannot be removed from the platform the way cm-7's interpreters partly can be — what is blocked is a slice of unauthorized agents, and disguised names plus user-level execution leave a bounded remainder the clause does not reach.
- T1543.001partialresponds — Malware repair software can quarantine or remove malicious .plist files and their associated payloads if they are identified as malware, but this does not eradicate living-off-the-land persistence mechanisms that evade signature or heuristic detection.
- T1543.002partialdetects — A.8.7's guidance (d) explicitly covers automated validation detecting unapproved files and unauthorized amendments, and modern EDR/file-integrity monitoring configured for Linux can surface creation or modification of `.service` unit files in `/etc/systemd/system` or user-level paths; however, the clause does not mandate systemd-specific monitoring, and basic signature-based AV implementations — which the clause itself notes are 'not usually adequate' — will not flag a `.service` file as malicious, so detection is a slice of well-instrumented implementations rather than a bounded remainder of a mechanism that mostly works.
- T1543.003partialdetects — A.8.7's malware detection software (AV/EDR) scans dropped service executables and drivers and can monitor service creation events, catching many instances of T1543.003, but the technique explicitly incorporates evasion (masquerading, hidden services via sc sdset, BYOVD signed drivers, rootkits) that bypass signature and behavioral detection, leaving a bounded but real remainder.
- T1543.003partialprevents — A.8.7's allowlisting (a), vulnerability reduction (c), and validation (d) can prevent unauthorized service executables from running, but Windows Services are a required OS mechanism that cannot be removed, and the technique has multiple bypass paths — modifying existing services via Registry, using inherently-allowed system utilities (sc.exe), and BYOVD with signed drivers — so what is prevented is a slice rather than the bulk.
- T1543.003partialresponds — A.8.7's malware detection and repair software (clause f) can detect a malicious Windows service created for persistence and remove or quarantine its payload, which is a containment/eradication action against a technique already underway; however, this response capability is secondary to the control's prevention/detection focus, and sophisticated variants — hidden services, BYOVD drivers, kernel-level rootkits — can evade both detection and removal, leaving a named remainder rather than a closed loop.
- T1543.004partialdetects — A.8.7's guidance (a) on detecting unauthorized software and (d) on automated validation of software/data content and investigating unapproved files could surface an unauthorized or modified plist in /Library/LaunchDaemons/, but the clause is general malware protection rather than mandating Launch Daemon-specific monitoring, and detection depends on whether the implementation actually instruments those directories and whether the plist is disguised as legitimate software.
- T1543.004partialprevents — A.8.7's application allowlisting (a), regular validation of software and data content (d), and malware detection (f) can prevent or catch unauthorized Launch Daemon plist files and their referenced payloads, but the technique requires elevated privileges to install — a domain belonging to A.8.2 — and Launch Daemons are a legitimate system mechanism, so the control addresses a slice rather than the bulk of the technique.
- T1543.004partialresponds — Malware repair software can quarantine or remove the malicious executables executed by the Launch Daemon, but may not remediate the modified plist files or custom scripts if they are not recognized as known malware signatures.
- T1543.005partialdetects — A.8.7 mandates malware detection and repair software with regular scanning of files, storage media, and network traffic, which can surface malicious container service modifications as anomalous files or unauthorized changes, but the control's scope is malware-focused and does not guarantee detection of all container service persistence mechanisms, especially those using legitimate tools like Docker or Kubernetes without dropping malware artifacts.
- T1546partialdetects — A.8.7's malware detection software (f) can detect payloads executed via event triggers when signatures match, and clauses (a) allowlisting and (d) validation of software/data content can surface unauthorized trigger mechanisms or amendments, but the trigger infrastructure itself (WMI subscriptions, cron jobs, registry Run keys, cloud event handlers) is system configuration that traditional AV frequently does not flag, leaving a substantial slice of the technique undetected.
- T1546partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can block unauthorized payloads and detect malicious content delivered via event triggers, but T1546 fundamentally abuses legitimate system mechanisms (WMI subscriptions, cron, launchd, cloud event handlers) that A.8.7 does not restrict, so the prevention reaches only a slice of the technique's surface.
- T1546partialresponds — A.8.7's malware repair capabilities can quarantine the malicious payloads invoked by the event triggers, but will likely leave the configuration changes (the triggers themselves) intact and may not address fileless or legitimate-binary payloads.
- T1546.001partialdetects — A.8.7's malware detection can catch the malicious payload when a file is opened via the modified association, and application allowlisting (guidance a) can detect execution of an unauthorized program triggered by the changed handler, but the core persistence mechanism — the registry modification itself — is not what A.8.7's primary scanning and detection mechanisms are designed to surface.
- T1546.001partialresponds — Malware repair software can quarantine the malicious executable referenced by the modified file association, breaking the persistence mechanism, though it may not always revert the registry change itself.
- T1546.002partialdetects — A.8.7's malware detection software and application allowlisting would catch the malicious .scr file (PE) when scanned or when an unauthorized executable attempts to run, but the persistence mechanism itself — registry modifications to SCRNSAVE.exe, ScreenSaveActive, ScreenSaverIsSecure and ScreenSaveTimeout — sits outside A.8.7's scope and would be surfaced by monitoring controls (A.8.16) rather than this one.
- T1546.002partialprevents — A.8.7's application allowlisting (a) and vulnerability reduction (c) can block the malicious .scr file from being placed or executed, but the technique's reliance on registry key manipulation means it can still succeed if the adversary has sufficient privileges to modify those keys, leaving a bounded remainder.
- T1546.002partialresponds — Malware repair software can quarantine the malicious screensaver executable to eradicate the payload, but it does not revert the registry modifications that constitute the persistence mechanism itself.
- T1546.003partialdetects — A.8.7's detection mechanisms — malware detection software (f), detection of unauthorized software (a), and automated validation of system content including unauthorized amendments (d) — can surface WMI event subscription creation through EDR/Sysmon-style telemetry on the WMI repository, but the clause does not mandate the specific instrumentation depth needed to reliably catch subscriptions created via legitimate Windows binaries (mofcomp.exe, WmiPrvSe.exe), and the clause itself states detection software alone is 'not usually adequate,' leaving a named remainder of evasion-capable variants.
- T1546.003partialresponds — Anti-malware and repair software can eradicate known malicious WMI event subscriptions and associated MOF files, though abuse of this native OS feature often requires manual remediation.
- T1546.004partialdetects — A.8.7's malware detection software can surface known malicious payloads embedded in shell config files, and clause (d)'s automated validation of software/data content for unauthorized amendments can catch modifications to /etc/profile, ~/.bashrc and similar files; however, the technique is specifically designed to evade detection by using legitimate shell mechanisms with custom or unknown commands that signature-based AV does not match, and file integrity monitoring is not mandated by the clause.
- T1546.004partialresponds — Anti-malware repair software can quarantine malicious binaries executed by the modified shell configuration or remove known malicious scripts, but may not revert manual text modifications of legitimate user configuration files.
- T1546.005partialprevents — A.8.7's malware detection (item f) may catch the payload that the trap executes, and regular validation of system content (item d) could surface unauthorized amendments to shell profiles, but the trap command itself is a built-in shell feature that cannot be removed by allowlisting — the shell is required, and the persistence mechanism is configuration rather than delivered malware, so what is prevented is a slice of the technique's instances rather than the technique.
- T1546.005partialresponds — Malware repair software can quarantine malicious scripts and kill offending shell processes that use the trap command, but it may not revert malicious trap configurations injected into legitimate user profiles.
- T1546.006partialdetects — A.8.7's guidance (d) explicitly calls for regular automated validation of software content and investigation of unauthorized amendments, which is the mechanism that would surface a modified Mach-O header, and (f) malware detection software can flag the tainted dylib if it carries a signature; however, the technique is designed to evade these controls (custom dylibs lack signatures, and removing LC_CODE_SIGNATURE defeats signature checks on the host binary), so detection depends on implementation choices such as file-integrity monitoring depth and allowlist coverage rather than being inherent to the clause.
- T1546.006partialprevents — A.8.7 includes software integrity validation (8.7d — investigating unauthorized amendments) and application allowlisting with hash checking (8.7a) that can block modified Mach-O binaries, but the technique specifically targets already-allowed binaries and removes LC_CODE_SIGNATURE to bypass signature checks, so prevention depends on implementation depth (e.g., hash-based allowlisting, continuous integrity monitoring) rather than being inherent to the clause.
- T1546.006partialresponds — Malware repair software can quarantine the modified binary or the malicious dylib, eradicating the persistence mechanism, with the remainder being non-malware system changes that require manual restoration.
- T1546.007partialdetects — A.8.7 mandates malware detection and repair software (clause f) that scans files and data content, which would surface the malicious helper DLL on disk and its execution when netsh.exe runs; however, the control's scope is broad and explicitly states detection software alone is 'not usually adequate', and it does not mandate the registry-monitoring depth needed to catch all variants of this persistence mechanism, leaving a bounded remainder where sophisticated or signed payloads evade signature-based detection.
- T1546.007partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent the malicious helper DLL from executing or detect its presence, but the technique leverages the legitimate netsh.exe binary to load the payload, the control is general guidance spanning many sub-mechanisms not all of which address registry-based persistence, and a conformant implementation focused on detection rather than strict allowlisting leaves the technique reachable.
- T1546.007partialresponds — Malware repair software can quarantine the malicious DLL, eradicating the payload, but it does not typically revert the registry modifications that established the persistence.
- T1546.008partialdetects — A.8.7(d) requires automated validation of software content and investigation of unauthorized amendments, which can detect binary replacement of accessibility features, and (a) covers detection of unauthorized software; however, the technique often uses legitimate Windows binaries (e.g., cmd.exe) that evade malware detection (f), and the IFEO debugger variant avoids binary replacement entirely, so detection is a slice rather than comprehensive coverage.
- T1546.008partialresponds — Malware repair software can quarantine malicious binaries dropped to replace accessibility features, but this eradication does not extend to the use of legitimate system binaries like cmd.exe.
- T1546.009partialdetects — A.8.7's malware detection and repair software (clause f) can detect the malicious DLL component when written to disk or scanned, and automated validation (clause d) can surface unauthorized files, but AppCert DLLs is a persistence mechanism that loads attacker DLLs into legitimate processes via a registry key — the registry modification and the DLL-load-into-foreign-process behavior are not what A.8.7's signature-oriented scanning is built to catch, and the technique is designed to evade traditional AV by abusing trusted API call paths.
- T1546.009partialprevents — A.8.7's malware detection and repair software (clause f) can catch the malicious DLL before it is placed, and application allowlisting (clause a) constrains which DLLs may reside on the system, but the technique's core mechanism — a Registry value under HKLM\System\CurrentControlSet\Control\Session Manager that causes the OS itself to load a DLL into every process calling CreateProcess-family APIs — is not directly addressed by A.8.7, and signed or otherwise clean-looking DLLs evade the slice the control does cover.
- T1546.009partialresponds — Malware repair software can quarantine or remove the malicious DLL used for persistence, though it may not revert the registry changes that point to it.
- T1546.010partialprevents — A.8.7's application allowlisting (a) and vulnerability reduction (c) can prevent the technique by blocking the malicious DLL or removing the registry path, but the control's malware-scanning core (f) acts after the DLL is placed and does not stop the initial write, so prevention is a slice rather than the bulk.
- T1546.010partialresponds — Malware repair software can quarantine the malicious DLL and remove the registry modification if the payload is identified as malware, eradicating the persistence mechanism.
- T1546.011partialdetects — A.8.7's malware detection software (f), automated validation of software/data content (d), and allowlisting (a) can surface application shimming via sdbinst.exe execution, new .sdb files in AppPatch directories, and registry modifications, but the technique abuses legitimate Windows infrastructure so detection depends on tool sophistication and is not comprehensive under the clause's general requirements.
- T1546.011partialprevents — A.8.7's application allowlisting (a), malware detection (f), and automated validation of software content (d) provide preventive mechanisms that can block unauthorized shim database creation or detect malicious sdbinst.exe usage, but the technique abuses legitimate Windows Application Compatibility Infrastructure where sdbinst.exe is a signed Microsoft binary, custom shims evade signature-based detection, and the control explicitly states malware detection software alone is not usually adequate — so prevention is a slice of the technique's surface, not the bulk.
- T1546.011partialresponds — Malware repair software can automatically quarantine or delete malicious shim databases (.sdb files) if they are identified as malware, providing automated eradication for known threats.
- T1546.012partialdetects — A.8.7's malware detection software (f) and application allowlisting (a) can detect the malicious debugger executable when written or launched, and clause (d)'s validation of software/data content and unapproved amendments can surface unauthorized registry changes, but the IFEO registry modification itself sits outside A.8.7's primary file-scanning and allowlisting mechanisms — registry-level monitoring belongs to A.8.16 — so what is caught is the payload and some unauthorized changes, not the persistence mechanism as a whole.
- T1546.012partialprevents — A.8.7's application allowlisting guidance (a) prevents IFEO injection when the adversary registers an unauthorized debugger binary, but the technique commonly abuses legitimate Windows binaries (cmd.exe, utilman.exe, etc.) that allowlisting permits, and the control's malware-scanning focus (f) addresses files rather than the registry-based persistence mechanism that is IFEO's actual vector.
- T1546.012partialresponds — Malware detection and repair software can quarantine malicious executables configured as IFEO debuggers, but this eradication does not extend to the abuse of legitimate system binaries like cmd.exe.
- T1546.013partialdetects — A.8.7's malware detection software (clause f) can flag malicious content in PowerShell profiles if signatures match, and clause (d)'s automated validation of software/data content and investigation of unapproved files or unauthorized amendments can surface profile modifications, but detection is not guaranteed — many profile-based attacks use legitimate PowerShell constructs that evade signature-based scanning, and coverage of profile directories depends on implementation choices.
- T1546.013partialprevents — A.8.7's malware detection software (f) can scan profile.ps1 files for malicious content and block execution, and application allowlisting (a) plus regular validation of software content (d) can prevent or catch unauthorized modifications to profile files; however, the profile is a legitimate PowerShell feature, sophisticated payloads can evade signature-based detection, and allowlisting does not necessarily cover modifications to existing trusted profile files, so prevention is partial rather than complete.
- T1546.013partialresponds — Malware repair software can quarantine a modified PowerShell profile if it contains known malicious signatures, but modifications using legitimate system commands or living-off-the-land techniques will often evade classification and removal.
- T1546.014partialdetects — A.8.7's clause (d) on investigating unapproved files and unauthorized amendments directly applies to a rule file dropped into /etc/emond.d/rules/, and behavioral EDR under (f)/(g) can surface emond spawning unexpected processes, but the technique is a persistence mechanism rather than a malware payload, so signature-based detection in (f) is unlikely to flag the plist rule itself and the (d) clause is scoped to systems supporting critical business processes rather than universally mandated.
- T1546.014partialprevents — A.8.7's application allowlisting (a), file/content validation (d), and malware detection (f) can partially prevent placement of malicious emond rule files, but emond is a built-in macOS daemon that cannot be removed (the platform requires it), and the technique abuses legitimate system functionality via a configuration file rather than installing foreign malware, so what is prevented is a slice with a bounded remainder.
- T1546.014partialresponds — Malware detection and repair software can quarantine known malicious payloads or plist files used for emond persistence, eradicating the foothold, though text-based configuration files may evade signature-based detection.
- T1546.015partialdetects — A.8.7's malware detection software (item f) can catch some COM hijacking payloads when they execute or are scanned, and item (d)'s automated validation of software/data content may surface unauthorized Registry amendments, but COM hijacking is specifically designed to be stealthy by hijacking references to legitimate components, and A.8.7 does not mandate Registry monitoring depth — implementations focused on file scanning alone miss the Registry-reference manipulation that is the technique's core mechanism.
- T1546.015partialprevents — A.8.7's application allowlisting (8.7a) and malware detection (8.7f) can prevent some COM hijacking payloads from executing, but the technique specifically abuses legitimate Windows COM infrastructure and Registry references that A.8.7 does not directly control; the hijacked Registry entry persists even when individual payloads are blocked, and the technique is designed to load through trusted system binaries (rundll32) or script: monikers that allowlisting often permits.
- T1546.015partialresponds — Malware repair software can quarantine the malicious payload referenced by the hijacked COM object, eradicating the adversary's code execution, though it may not restore the modified Registry keys to their legitimate state.
- T1546.016partialdetects — A.8.7 requires regular automated validation of software/content, malware scanning of received files/media/email/webpages, collection of new malware intel, and procedures to detect/report malware (including during maintenance), which surfaces malicious installer scripts or packages when they match known signatures or trigger scans; partial because it depends on up-to-date signatures, does not inherently detect novel/custom installer abuse, and leaves gaps for encrypted/evasive delivery or pre-compromise maintainer scripts outside scanned vectors.
- T1546.016partialprevents — A.8.7's application allowlisting (a), scanning of received data and email attachments (f.1, f.2), and validation of software content for unauthorized amendments (d) all act against installer-borne malware, but the technique specifically abuses LEGITIMATE installers with modified scripts that inherit elevated permissions, so allowlisting does not stop a trusted package and malware detection can be evaded by scripts that look like normal pre/post-install behavior; installers are also platform-required for legitimate software and cannot be removed wholesale, leaving a bounded remainder the control cannot reach.
- T1546.017partialdetects — A.8.7's malware detection software (clause f) can detect the malicious binary when udev triggers its execution, and clause (d)'s regular automated validation of system content could surface unauthorized changes to udev rule files, but the clause's primary mechanisms — signature-based scanning of files, email, and web content — are not aimed at persistence-via-legitimate-subsystem abuse, and the udev rule itself is a configuration file that looks legitimate to most malware scanners.
- T1546.017partialprevents — A.8.7's malware detection and application allowlisting can prevent the malicious payload triggered by the udev rule from executing, but the technique's core mechanism — creating/modifying udev rule files — requires root and operates at a layer (kernel device manager) that A.8.7's controls do not address; the rule is still planted and triggered even when the payload is blocked.
- T1546.017partialresponds — Malware repair software can quarantine the malicious payload executed by the udev rule and terminate its processes, disrupting the persistence mechanism, though it may not revert the unauthorized configuration changes to the rule files themselves.
- T1546.018partialdetects — A.8.7's malware detection software and its guidance on validating software/data content for unauthorized amendments can surface malicious .pth files or tampered sitecustomize.py/usercustomize.py, but the technique deliberately abuses legitimate Python mechanisms, so signature-based and heuristic detection catch only a slice of instances rather than the bulk.
- T1546.018partialprevents — A.8.7's allowlisting (a), file-content validation (d), and malware scanning (f) can catch some unauthorized .pth files or malicious sitecustomize.py content, but the technique abuses legitimate Python interpreter features that are typically required on systems, and the embedded code can be obfuscated or novel enough to evade signature-based detection, leaving a substantial remainder.
- T1546.018partialresponds — Malware repair software can quarantine or delete malicious Python scripts and configuration files used for persistence, eradicating the artifact if it is recognized as malware.
- T1547partialdetects — A.8.7's malware detection and repair software (clause f) and its automated validation of software and data content (clause d) are well-suited to detecting autostart persistence: AV/EDR products routinely flag modifications to Run keys, startup folders, and other boot/logon autostart locations as known malware behaviors, and integrity checks surface unauthorized amendments to system configuration. The remainder is the evasion surface — sophisticated adversaries using legitimate system tools, kernel-level modifications, or novel persistence mechanisms can evade signature and behavioral detection, and A.8.7 is not a comprehensive host-telemetry control like si-4.
- T1547partialprevents — A.8.7's application allowlisting (a), malware detection (f), and regular validation (d) directly prevent many T1547 instances by blocking unauthorized executables from running at startup, but the technique can leverage legitimate binaries (LOLbins), modify existing autostart entries, or use kernel-level mechanisms that allowlisting and detection do not catch — and the clause itself acknowledges that detection software alone is not adequate.
- T1547partialresponds — Malware detection and repair software can quarantine malicious files and remove their associated autostart entries, eradicating the foothold, but this only addresses persistence mechanisms that rely on identifiable malware rather than abused legitimate binaries.
- T1547.001partialprevents — A.8.7's application allowlisting (a) and malware detection (f) directly prevent unauthorized programs referenced in registry run keys or startup folders from executing, but the technique can bypass these via masquerading with legitimate-looking programs or by referencing already-allowlisted binaries, and the control is broader than a dedicated persistence-prevention mechanism.
- T1547.001partialresponds — A.8.7 mandates malware detection and repair software, which can eradicate the technique by removing malicious registry run keys or startup folder entries if they are identified as part of a malware infection.
- T1547.002partialdetects — A.8.7's malware detection software (item f) and application allowlisting (item a) can surface some instances of authentication package abuse — particularly known-malicious DLLs or unauthorized binaries flagged during scans — but the technique specifically abuses a trusted system mechanism (LSA loading) designed to evade signature-based detection, and A.8.7 does not mandate registry-monitoring depth that would catch the HKLM\...\Lsa\Authentication Packages modification as a primary indicator.
- T1547.002partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent an unauthorized DLL from being loaded by LSA, but the technique abuses a legitimate Windows mechanism that must function for authentication to work, so the preventive measure has a bounded remainder analogous to cm-7 vs T1059: the allowlist must permit LSA's DLL loading, and detection only catches known malware, leaving a slice rather than full coverage.
- T1547.002partialresponds — A.8.7's malware detection and repair software can detect the malicious authentication package DLL once loaded and quarantine or remove it, which is eradication once the persistence technique is underway; however, the control's response is limited to automatic removal by the detection tool and does not extend to broader incident response activities such as investigating the registry modification, checking for additional persistence mechanisms, or formal incident handling, which is what A.5.26 covers.
- T1547.003partialdetects — A.8.7's malware detection and repair software (8.7f) can scan and detect the malicious DLL if it carries known signatures, and 8.7d's investigation of unauthorized amendments can surface the new W32Time\TimeProviders subkey, but the technique specifically abuses a legitimate Windows service to achieve persistence, the DLL may be unknown, signed, or fileless, and A.8.7's guidance does not mandate the registry-change telemetry or behavioral detection needed to reliably catch abuse of a trusted mechanism.
- T1547.003partialprevents — A.8.7's malware detection (f) and application allowlisting (a) can catch or block the malicious DLL before it is loaded as a time provider, but the technique abuses legitimate Windows functionality (W32Time loading a registered DLL) and the persistence mechanism (registry subkey creation under TimeProviders) is not directly addressed by malware protection controls; admin privilege requirement is an access-control matter outside A.8.7's scope.
- T1547.003partialresponds — Malware repair software can quarantine the malicious DLL used for the time provider, eradicating the payload, though it may not revert the specific registry modifications.
- T1547.004partialprevents — A.8.7's allowlisting (a), email/attachment scanning (f.2), and AV detection can prevent the malicious DLL from arriving on the system or block its execution, but the technique abuses a trusted Windows component (Winlogon) whose helper DLLs are loaded by a system process that allowlists typically permit, and A.8.7 does not address the registry modification that establishes the persistence — so what is prevented is the delivery/detection slice, not the abuse mechanism itself.
- T1547.004partialresponds — Malware repair software can quarantine and remove the malicious DLLs or executables loaded via Winlogon, neutralizing the threat, though it may not revert the modified registry keys themselves.
- T1547.005partialdetects — A.8.7's malware detection and repair software (clause f) can surface SSP abuse when the loaded DLL matches a known signature or exhibits detectable behaviour, and clause (d)'s automated validation of software content can flag unauthorized registry amendments, but the technique is specifically designed to persist stealthily across reboots and often evades signature-based detection when adversaries use legitimate or custom DLLs, so what is caught is a slice rather than the bulk.
- T1547.005partialprevents — A.8.7's allowlisting (a), malware detection (f), and integrity validation (d) can prevent some SSP-abuse instances by blocking unauthorized DLLs or detecting registry modifications, but the technique loads malicious code through a trusted system process (LSA) using custom DLLs that evade signature-based detection and abuse legitimate OS functionality the control does not specifically address.
- T1547.005partialresponds — Anti-malware software can quarantine the malicious DLL used for persistence, eradicating the payload, though it may not revert the registry modifications used to register the Security Support Provider.
- T1547.006partialdetects — A.8.7's detection mechanisms — malware detection software, application allowlisting (8.7a), regular automated validation of software content (8.7d), and signature verification on macOS kexts — can surface unauthorized kernel module loading, but the technique runs at Ring 0 and is specifically designed to evade the very tools the clause mandates, so detection is real but bounded.
- T1547.006partialprevents — A.8.7's allowlisting (a), defense-in-depth (g), and vulnerability reduction (c) guidance can be extended to restrict unauthorized kernel module loading, but the control's primary mechanism—malware detection and repair software (f)—is largely ineffective against kernel-level rootkits operating at Ring 0 below traditional AV visibility, and the clause does not specifically mandate kernel integrity monitoring, secure boot, or module-signing enforcement.
- T1547.006partialresponds — Malware protection includes repair software and environment isolation that can quarantine malicious kernel module files and contain the host, though rootkits running at Ring 0 often evade in-memory eradication and require deeper system restoration.
- T1547.007partialdetects — A.8.7's detection mechanisms — malware detection software (f), detection of unauthorized software (a), and validation of software/data content including investigation of unauthorized amendments (d) — can surface the malicious payload or the plist modification, but the technique is a persistence mechanism via configuration change rather than malware delivery, and A.8.7 does not mandate file-integrity or plist-specific monitoring, so what is caught depends on implementation choices.
- T1547.007partialprevents — A.8.7's mechanisms partially prevent this macOS persistence technique: application allowlisting (8.7a) can block the malicious app from executing, validation of software/data content and investigation of unauthorized amendments (8.7d) can surface the plist modification, and malware detection (8.7f) can catch the payload. However, A.8.7 is a malware control, not a persistence control, and does not specifically target macOS login item plist files; gaps remain where the malicious app is allowlisted (e.g., a legitimate binary abused via plist reference) or where the plist modification occurs outside the monitored scope.
- T1547.007partialresponds — A.8.7 includes malware repair software that can quarantine malicious payloads and remove associated persistence entries like modified plist files, but this eradication only occurs if the payload is recognized as known malware.
- T1547.008partialdetects — A.8.7 mandates malware detection and repair software with regular scanning of files, email, and web content, which can surface a malicious LSASS driver on disk or in transit, but the control's guidance does not guarantee runtime detection of the driver's execution within the LSASS process, leaving a nameable gap.
- T1547.008partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can partially prevent unauthorized LSASS drivers by blocking the tools that place them and detecting known malicious DLLs, but LSASS drivers are loaded by the OS itself from system locations where standard allowlisting and AV often have limited visibility, and signed or sophisticated drivers can evade detection.
- T1547.008partialresponds — Malware repair software can quarantine the malicious DLL used as an LSASS driver, eradicating the file-based component of the persistence mechanism, though it may not revert the associated registry modifications.
- T1547.009partialdetects — A.8.7 requires regular automated scanning of software and data content, which can surface unauthorized shortcut modifications on disk, but the control's detection scope is bounded by scan frequency and the adversary's ability to place the shortcut where scanning is not configured to look.
- T1547.009partialprevents — A.8.7's application allowlisting (a) can prevent unauthorized payloads referenced by malicious shortcuts from executing, and its automated validation of software/data content (d) can surface unapproved files in startup locations, but shortcuts are a legitimate file format and the technique can abuse existing legitimate binaries or modify existing shortcuts, which the control's general malware-scanning posture does not specifically target.
- T1547.009partialresponds — Malware repair software can quarantine and remove malicious shortcut (LNK) files used for persistence, but may not remediate legitimate shortcuts that have been modified to include malicious command-line arguments.
- T1547.010partialprevents — A.8.7's malware detection (8.7f) and application allowlisting (8.7a) can catch known-malicious DLLs before they are written to System32, but the technique uses a legitimate Windows mechanism (print spooler loading a monitor DLL) and a custom or repacked DLL evades signature-based detection, so the preventive coverage is a slice rather than the bulk.
- T1547.012partialdetects — A.8.7's malware detection software (f) and automated validation of software/data content (d) can surface unauthorized additions to the print processor directory and registry key modifications, and AV may flag the malicious DLL itself; however, the technique abuses a legitimate Windows mechanism, so detection depends on signature coverage of the specific DLL and on whether the implementation monitors the print processor registry path and directory — neither of which A.8.7 mandates, leaving a bounded but real remainder.
- T1547.012partialprevents — A.8.7's application allowlisting (8.7a) and regular validation of software content (8.7d) can prevent unauthorized DLLs from being placed in the print processor directory or detect them, but the technique abuses a legitimate Windows service (spoolsv.exe) using legitimate APIs (AddPrintProcessor), and the spooler service itself must run for printing to function — so the underlying mechanism cannot be removed, and a custom DLL may evade signature-based detection, leaving a meaningful remainder the control does not reach.
- T1547.012partialresponds — A.8.7's malware repair software can quarantine or remove the malicious DLL used as the print processor, eradicating the payload, though it may leave the registry modifications intact.
- T1547.013partialdetects — A.8.7(d) explicitly requires investigating unapproved files and unauthorized amendments, and A.8.7(f) malware detection can scan .desktop files, so the control has mechanisms that surface unauthorized XDG Autostart entries; but detection depends on whether the autostart directories fall within the validation scope and whether the AV product inspects Linux .desktop files, leaving a bounded remainder.
- T1547.013partialresponds — Malware repair software can quarantine malicious binaries or .desktop files used for autostart persistence, but provides only partial eradication because it may not identify or remove entries that abuse legitimate system tools.
- T1547.014partialdetects — A.8.7's malware detection software (clause f) can detect the malicious payload referenced by StubPath if it carries known signatures, and clause (d)'s automated validation of software/data content can surface unauthorized Registry amendments, but the clause sets scope rather than mandating Registry-monitoring depth: an implementation scanning only files and processes is fully conformant and sees nothing of the Active Setup key creation itself, so what is caught is a slice the implementer chose rather than a bounded remainder of a mechanism that mostly works.
- T1547.014partialprevents — A.8.7's application allowlisting (clause a) and malware detection (clause f) can prevent the unauthorized binary referenced by StubPath from executing, but the control does not address Registry key creation itself, and Active Setup is a built-in Windows mechanism that can be abused via LOLBins or masquerading where allowlisting provides no barrier.
- T1547.014partialresponds — Malware repair software can quarantine malicious payloads and remove their associated persistence mechanisms, but this eradication only applies when the activity is identified as malware, leaving manual or non-malware registry modifications unaddressed.
- T1547.015partialdetects — A.8.7's malware detection software (clause f) and rules against unauthorized software (clause a) can detect login item persistence if the endpoint protection monitors macOS-specific persistence locations (backgrounditems.btm, SMLoginItemSetEnabled API calls) and scans the executables launched at login, but the control does not specifically mandate monitoring of login items and detection depends on the tool's macOS coverage.
- T1547.015partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent some login item persistence by blocking unauthorized executables from running, but the technique adds login items via AppleScript or the SMLoginItemSetEnabled API using built-in macOS tooling, which allowlisting does not necessarily stop, and the clause is general-purpose rather than targeting this specific persistence mechanism.
- T1547.015partialresponds — Malware repair software can quarantine malicious executables and remove their associated login items once detected, but this eradication only applies to recognized malware and leaves behind persistence mechanisms created using native tools or unrecognized scripts.
- T1548partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, which can surface known malware used to abuse elevation control mechanisms, but the control's scanning scope does not inherently cover all abuse methods such as UAC bypasses or sudo abuse that do not involve malicious files.
- T1548.001partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent the variant where adversaries set setuid on their own malware, but the technique also encompasses abusing legitimate setuid/setgid binaries already on the system, which A.8.7 does not address — that is a configuration/vulnerability concern outside its scope.
- T1548.002partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can surface known UAC bypass malware samples, but the control's scanning scope does not inherently detect the behavioral technique of UAC bypass itself, especially when executed via trusted binaries or fileless methods.
- T1548.004partialdetects — A.8.7's malware detection software and application allowlisting can catch known malicious programs calling AuthorizationExecuteWithPrivileges, but the technique is specifically designed to evade these controls by using modified legitimate programs that already pass allowlisting and by leveraging a deprecated but functional API that performs no origin or integrity validation, so a meaningful slice of instances goes undetected.
- T1548.004partialprevents — A.8.7's application allowlisting (a) prevents unauthorized programs from running and thus from calling AuthorizationExecuteWithPrivileges, and user awareness/training (m) reduces the rate at which users enter credentials for unexpected prompts; however, the technique explicitly works by modifying legitimate programs already on the system (which pass allowlisting) and trained users still sometimes comply, so what is prevented is a slice rather than the bulk.
- T1548.006partialdetects — A.8.7's malware detection software can identify the malicious binary or AppleScript payload used to carry out TCC manipulation, but it does not specifically detect the TCC database modification, environment variable abuse (e.g., loading an adversary-controlled TCC.db), or exploitation of legitimate applications with existing permissions — which is the core mechanism of T1548.006.
- T1548.006partialprevents — A.8.7's application allowlisting (8.7a) and malware detection can block some vectors where a malicious binary must execute, but TCC manipulation primarily abuses already-trusted macOS applications and manipulates the TCC database itself — a privilege-escalation/abuse-of-trusted-mechanism technique that malware-focused controls do not directly address.
- T1548.006partialresponds — Malware repair software can quarantine the malicious executables or scripts attempting to manipulate TCC, eradicating the offending process, though it does not revert unauthorized permissions already written to the TCC database.
- T1552.006partialdetects — A.8.7 mandates malware detection and repair software that scans files on electronic storage media and systems, which would surface a malicious XML file or script placed in SYSVOL during automated validation, but the control's scope is malware detection and not credential harvesting from legitimate administrative artifacts, so what is caught is the delivery mechanism rather than the technique itself.
- T1553partialdetects — A.8.7's scanning (8.7f) and application allowlisting (8.7a) can detect some manifestations of T1553 — unauthorized software executing despite trust controls, malware delivered via subverted signing, and registry modifications flagged by EDR-style tools — but the technique's core act of subverting trust mechanisms (stealing certificates, modifying Mark-of-the-Web, disabling SmartScreen) is not what A.8.7's malware-detection mechanisms are designed to surface, leaving a bounded remainder.
- T1553.001partialdetects — A.8.7's malware detection software and file-integrity validation (clause d) can surface the resulting malicious payload and unauthorized amendments to application bundles, but Gatekeeper bypass is specifically engineered to evade these controls — manipulating quarantine xattrs, exploiting logic errors, and using delivery paths (USB, network shares, curl) that bypass the scan points named in clause f — so the bypass mechanism itself largely escapes detection.
- T1553.001partialprevents — A.8.7's general malware protection measures (application allowlisting per 8.7a, malware scanning per 8.7f, blocklisting per 8.7b) would catch some malicious payloads attempting to execute via Gatekeeper bypass, but the technique is specifically designed to subvert macOS's built-in trust mechanism, and A.8.7 does not specifically address Gatekeeper or the file-attribute manipulation the technique exploits, leaving a substantial remainder where the bypass succeeds despite the control.
- T1553.001partialresponds — A.8.7's malware repair software quarantines and removes the malicious payload that bypassed Gatekeeper, eradicating the immediate threat, though it does not restore the subverted quarantine flags or patch the underlying bypass vulnerability.
- T1553.003partialdetects — A.8.7's detection mechanisms (malware scanning, application allowlisting per item a, and investigation of unauthorized amendments per item d) can surface some artifacts of SIP/trust provider hijacking — known-malicious replacement DLLs, unauthorized Registry modifications, and unapproved DLL loads — but the technique is specifically engineered to subvert the signature-validation chain that A.8.7's malware detection software fundamentally relies on, and the Registry-level tampering that enables the hijack is not a primary monitoring target of the clause.
- T1553.003partialprevents — A.8.7's application allowlisting (8.7a) and change management aspects can prevent unauthorized DLLs from running and block unauthorized Registry modifications that the technique relies on, but the control's primary mechanism (malware detection software, 8.7f) depends on the signature/trust validation chain that T1553.003 specifically subverts, and the technique can hijack existing DLLs to bypass allowlisting, so prevention is partial rather than complete.
- T1553.004partialdetects — A.8.7 mandates malware detection and repair software that scans files and storage media, which can surface a malicious root certificate on disk, but the control's scanning scope is file-based and does not guarantee detection of a certificate installed directly into the system trust store without a file artifact or via memory-only techniques.
- T1553.004partialresponds — Malware repair software can eradicate malicious root certificates when they are deployed by or associated with known malware, but it does not remove certificates installed manually by an adversary during an intrusion.
- T1553.005partialdetects — A.8.7's malware detection and repair software (clause f) scans downloads and container files, and clause (g) explicitly accounts for evasive techniques like encrypted containers, so the malicious payload inside .iso/.vhd/.gzip archives is detectable when AV signatures or behavioral rules match; however, the MOTW bypass itself — the abuse of NTFS ADS inheritance across container formats — is a Windows trust-control subversion that A.8.7 does not specifically instrument for, and archive scanning depth, encrypted containers, and unknown payloads leave a bounded remainder.
- T1553.005partialprevents — A.8.7's scanning (f), allowlisting (a), and blocklisting (b) intervene at delivery and execution of the container-borne payload, but the technique specifically targets MOTW — a Windows feature A.8.7 does not address — and container formats can be crafted so scanners do not deeply inspect contents, leaving a meaningful slice of instances reaching execution.
- T1553.005partialresponds — A.8.7's malware repair and isolation functions can quarantine the container files or extracted payloads used in the bypass, eradicating the artifacts provided they are recognized as malicious.
- T1554partialprevents — A.8.7(d) explicitly calls for regular automated validation of software content and investigation of unauthorized amendments, and (a) application allowlisting with hash/signature checks can block execution of tampered binaries, both directly addressing T1554's binary modification; however, the technique is designed to preserve normal execution flow (IAT hooking, entry-point patching) and evade signature-based detection, so the control's mechanisms catch a slice rather than fully preventing the technique.
- T1554partialresponds — A.8.7 deploys malware repair software that can quarantine or remove infected host binaries, effectively eradicating the adversary's persistence mechanism once discovered.
- T1555partialdetects — A.8.7 mandates malware detection and repair software that scans files and data content, which can surface credential-dumping tools when they are written to disk or transferred, but the control's mechanism is file-scanning and not runtime behaviour monitoring, so in-memory credential access and legitimate-tool abuse pass unseen.
- T1555.001partialdetects — A.8.7 mandates malware detection and repair software that scans files and storage media, which can surface credential-dumping tools when they are written to disk or executed, but the control's scanning scope does not inherently reach in-memory credential access or the legitimate security utility used to perform it.
- T1555.003partialdetects — A.8.7's malware detection and repair software can flag known credential-stealing tools (e.g., mimikittenz, browser credential dumpers) and anomalous access to browser credential stores, but the technique can be executed with built-in system utilities (SQL queries, CryptUnprotectData, process memory reads) that are not malware and would not be caught by anti-malware signatures, leaving a substantial slice undetected.
- T1555.003partialprevents — A.8.7's malware detection, application allowlisting, and vulnerability reduction can block some tools used to extract browser credentials, but the technique can be executed with built-in OS utilities (SQL queries, CryptUnprotectData) or with malware that evades detection, and A.8.7 does not address credential-store protection directly.
- T1555.003partialresponds — A.8.7 includes malware repair software that quarantines and removes malicious files and processes, which eradicates the credential-stealing malware often used to execute this technique, though it does not address manual or script-based extraction by an adversary.
- T1555.004partialdetects — A.8.7's malware detection and repair software can detect credential manager access when known credential-dumping malware (e.g., Mimikatz) is involved, and EDR-style tools may flag suspicious access patterns to vault files, but the technique can be executed entirely with legitimate Windows utilities (vaultcmd.exe, rundll32.exe keymgr.dll, CredEnumerate API) that A.8.7's malware-focused controls are not designed to flag, leaving a substantial undetected remainder.
- T1555.004partialprevents — A.8.7's malware detection and application allowlisting (8.7a, 8.7f) can prevent malware-based credential dumping tools (e.g., Mimikatz) from running, but the technique can equally use legitimate Windows utilities (vaultcmd.exe, CredEnumerate API, rundll32.exe keymgr.dll) that A.8.7 does not block, and the control's scope is malware rather than credential-store protection.
- T1555.005partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, which can surface credential-dumping tools when they are written to disk or traverse a network gateway, but the technique's in-memory extraction and abuse of a legitimate password manager's own process evade file-based and gateway scanning, leaving a substantial observable gap.
- T1556partialdetects — A.8.7's malware detection and application allowlisting can surface T1556 when the authentication modification is delivered via malware (e.g., a malicious DLL loaded into LSASS, credential-theft tooling recognized as malicious), but the technique is fundamentally about modifying authentication processes — direct PAM/SAM edits, legitimate-tool abuse, or configuration tampering fall outside A.8.7's malware-centric detection scope.
- T1556partialprevents — A.8.7's application allowlisting (8.7a) and malware detection (8.7f) can prevent some T1556 instances where malware is used to modify authentication processes (e.g., malicious PAM modules, credential-stealing payloads), but the technique is broader than malware — adversaries can modify authentication via legitimate admin tools, registry edits, or built-in OS capabilities that allowlisting does not block and that malware scanners do not flag as malicious.
- T1556.001partialdetects — A.8.7's malware detection and repair software, scanning, and allowlisting can surface the Skeleton Key malware component (file artifacts, known signatures, unauthorized code execution), but the technique's core — in-memory patching of LSASS on a domain controller — is specifically designed to evade these mechanisms, and the clause itself states that malware detection software alone is not usually adequate.
- T1556.001partialprevents — A.8.7's malware detection, application allowlisting, and isolation of critical environments (domain controllers qualify) provide some preventive coverage against the Skeleton Key malware, but the technique operates in-memory by patching LSASS using legitimate system APIs, which the clause's own guidance (g) acknowledges as an evasive technique that malware protection must specifically contend with — so prevention is a slice, not a closure.
- T1556.001partialresponds — A.8.7 includes response mechanisms — malware detection and repair software (f), isolation of environments (k), and recovery planning (j) — that act once malware is present, but Skeleton Key is sophisticated in-memory malware on a domain controller that evades standard AV signatures and persists until reboot, so detection-and-repair reaches only a slice of this technique.
- T1556.002partialdetects — A.8.7's malware detection and repair software (item f) and unauthorized-software detection (item a) can surface a malicious password filter DLL where signatures or behavioral rules match, but the technique deliberately abuses a legitimate Windows mechanism with a custom DLL that often evades signature-based AV, and the clause does not mandate the LSA-configuration or behavioral monitoring that would reliably catch registration of the filter.
- T1556.002partialprevents — A.8.7's application allowlisting (8.7a) and malware scanning (8.7f) can prevent an unauthorized password filter DLL from being placed on the system or detect it before registration, but the technique requires administrative privileges to modify the LSA notification packages registry key and place a DLL in a system directory, and once an attacker holds those privileges they can often bypass allowlisting or disable scanning; the control is not specifically designed to protect authentication subsystem integrity, so what is prevented is a slice rather than the bulk.
- T1556.002partialresponds — Anti-malware repair software can quarantine or remove the malicious DLL, eradicating the file itself, but may not revert the registry changes used to register the filter or address the credentials already harvested.
- T1556.003partialdetects — A.8.7's clause (d) on automated validation of software content and investigation of unauthorized amendments, plus defense-in-depth placement (g), can surface tampering with PAM files and libraries, but the core mechanism — malware detection and repair software under (f) — generally does not catch trojanized authentication modules because the technique modifies legitimate system files rather than introducing recognizable malicious payloads, so detection is a slice rather than the bulk.
- T1556.003partialprevents — A.8.7 requires malware detection and repair software, application allowlisting, and vulnerability reduction, which together can prevent unauthorized modification of PAM modules, but the control's reliance on configuration and detection means it cannot guarantee prevention against all PAM tampering.
- T1556.003partialresponds — Malware repair software can quarantine or delete malicious PAM modules, eradicating the adversary's artifact, though restoring the legitimate module may require separate recovery actions.
- T1556.004partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and web content, and a patched OS image on a network device is a file that could be caught during integrity validation or scanning, but the control's primary detection mechanisms are not designed to find a hardcoded password in firmware, so detection is a possible but incidental slice.
- T1556.007partialdetects — A.8.7's malware detection software and application allowlisting can surface malicious DLLs injected into authentication processes (e.g., AzureADConnectAuthenticationAgentService) and unauthorized modifications to authentication service configurations, but the control is general-purpose malware protection rather than authentication-process monitoring, and sophisticated backdoors like MagicWeb are designed to evade standard endpoint malware detection.
- T1556.008partialdetects — A.8.7's malware detection and repair software (item f) scans files on computers and storage media, which would surface a malicious network provider DLL on disk if it matches a signature or sits in a scanned location, and item (d)'s automated validation of software content can flag unauthorized files; however, the technique is specifically designed to evade traditional detection by using a legitimate Windows mechanism (NPLogonNotify via mpnotify.exe) and a custom DLL may carry no signature, reside in a trusted system path, and only load during logon, so detection is a slice rather than a bounded remainder of a working mechanism.
- T1556.008partialprevents — A.8.7's application allowlisting (a) and malware detection/repair software (f, g) can prevent or detect the malicious network provider DLL before it registers as a credential manager, but the technique abuses a legitimate Windows authentication mechanism (NPLogonNotify) that A.8.7 does not specifically target, and the DLL's credential-capture function is not what signature-based or allowlist controls are designed to catch.
- T1556.008partialresponds — Malware repair software can quarantine and remove the malicious DLL once identified, halting the credential capture, though it may not revert the registry modifications used to register it.
- T1558.005partialdetects — A.8.7 mandates malware detection and repair software that scans files on disk and in transit, which can surface a ccache file being collected or exfiltrated by malware, but the control's scope is malware-based collection and not the credential-theft technique itself, so detection is a slice of the technique's observable footprint.
- T1559partialdetects — A.8.7's endpoint-resident malware detection tools (g.1) and application allowlisting (a) can surface anomalous IPC activity such as unexpected COM/DDE invocation or unauthorized processes initiating IPC, but the clause's primary mechanisms — file/email/web scanning for malware payloads (f) — do not address abuse of legitimate OS inter-process mechanisms, which is what T1559 actually does.
- T1559partialprevents — A.8.7's allowlisting (8.7a) and malware detection (8.7f) can prevent unauthorized payloads delivered via IPC, but the IPC mechanisms themselves (COM, DDE, sockets, pipes) are legitimate OS features that cannot be blocked without breaking the system, and many T1559 abuses leverage built-in functionality rather than introducing detectable malware.
- T1559.001partialdetects — A.8.7's malware detection and repair software can surface some COM abuse via behavioral indicators (anomalous child processes spawned via COM, suspicious script hosts invoking COM interfaces), but COM is a native, legitimate Windows IPC mechanism whose abuse is largely fileless and uses signed Microsoft binaries, so detection depends on the depth of behavioral telemetry the implementation chooses rather than being inherent to the clause's required mechanisms.
- T1559.001partialprevents — A.8.7's allowlisting (8.7a), blocklisting (8.7b), and vulnerability reduction (8.7c) can prevent unauthorized executables that invoke COM from running, but COM is a native Windows IPC mechanism required by the OS itself and cannot be removed or disabled without breaking the platform — so the technique's underlying mechanism stays reachable to any code that does execute, mirroring the cm-7 vs T1059 shape where the required interpreter cannot be taken away.
- T1559.001partialresponds — A.8.7 mandates malware repair software and isolation procedures that can contain and eradicate malicious processes abusing COM, though this response is limited to activities successfully identified as malware.
- T1559.002partialdetects — A.8.7 mandates malware detection and repair software that scans email attachments, downloads and webpages (f.1-3), and DDE-based payloads in Office documents and CSV files are well-known and broadly signature-detectable by modern AV; the bounded remainder is novel or obfuscated DDE payloads that evade signature-based detection and DDE invoked remotely via DCOM after initial compromise.
- T1559.002partialprevents — A.8.7's email/web scanning and blocklisting can prevent some DDE delivery vectors (malicious Office documents, hosted web content), but DDE is a legitimate Windows protocol that can be abused through legitimate Office applications and invoked remotely via DCOM, so the technique can still execute even with the control fully implemented.
- T1559.002partialresponds — Anti-malware repair software can quarantine poisoned documents and kill malicious processes spawned via DDE, though it may not contain purely fileless or living-off-the-land command execution.
- T1559.003partialdetects — A.8.7's application allowlisting (8.7a) and behavioral capabilities in modern malware detection tools can surface unauthorized XPC services or anomalous IPC patterns, but the technique's core vector — abusing legitimate XPC APIs in allowed applications to pass malicious content to a root daemon — is not addressed by A.8.7's primary file/email/web scanning mechanisms, leaving a substantial remainder undetected.
- T1559.003partialresponds — A.8.7's malware detection and repair software (clause f) and environment isolation (clause k) provide eradication and containment capability once the technique is underway, but XPC Services abuse exploits legitimate macOS inter-process communication and improper validation in trusted applications, so traditional malware detection is unlikely to surface the malicious content and the response reaches only what the detection layer can identify.
- T1560.001partialprevents — A.8.7's guidance item (a) on application allowlisting can prevent unauthorized third-party archive utilities (7-Zip, WinRAR, WinZip) from running, but the technique explicitly relies on pre-installed utilities (tar, zip, makecab, diantz, certutil, xcopy) that any functional allowlist must permit, so the built-in tools central to T1560.001 remain available.
- T1561partialdetects — A.8.7's malware detection software (item f) and unauthorized-software controls (item a) can detect known disk-wipe malware and unauthorized wiping tools, but the technique is frequently executed via legitimate system utilities (dd, sdelete, diskpart, cipher /w) or novel destructive payloads that signature-based and allowlist controls do not catch, leaving a substantial slice of the technique unobserved.
- T1561partialprevents — A.8.7's allowlisting (a), anti-malware detection (f), and defense-in-depth (g) can prevent the wiper malware from executing on endpoints, but disk wipe is also performed via legitimate tools (dd, erase, Network Device CLI) by adversaries with valid credentials or direct write access, which allowlisting will not block and anti-malware will not flag — so prevention is real but bounded.
- T1561partialresponds — A.8.7's malware detection and repair software (f) can detect disk-wiping malware in flight and quarantine/remove it, and clause (k) supports isolating environments to contain spread — genuine containment and eradication once the technique is underway. But A.8.7 is primarily a preventive control, the response capabilities are secondary to its scanning-before-use purpose, and the damage already realised (wiped disk sectors) is not addressed by response — that boundary belongs to A.8.13 `recovers`.
- T1561.001partialresponds — A.8.7's malware detection and repair software (f) and environment isolation (k) can act on wiping malware in flight by detecting the wiping process and containing its spread, but the control is primarily preventive and its response elements are a subset of its purpose; the disk content already overwritten is not addressed by these response actions.
- T1561.002partialdetects — A.8.7's malware detection and repair software (clause f) can detect the malware payload that performs disk structure wipe before it executes, and clause (d)'s automated validation of data content can surface unauthorized amendments to disk structures — but detection is bounded by evasion techniques (clause g.2 explicitly names them), zero-day wipers, and the destructive nature of the technique: once the MBR/partition table is overwritten the system is down and detection is forensic rather than actionable.
- T1561.002partialprevents — A.8.7's anti-malware software, allowlisting, and blocklisting can block the malware that delivers disk structure wipes before it executes, but the control itself states detection software alone is 'not usually adequate,' and the technique often runs after an adversary has already established a foothold with malware that evades detection or uses legitimate tools to overwrite MBR/partition tables.
- T1561.002partialresponds — A.8.7 includes procedures and responsibilities for dealing with malware attacks and recovering from them, which engages the response process once the wipe is underway, but the control's primary mechanism is prevention and detection, not containment and eradication of an already-executing disk wipe.
- T1564partialdetects — A.8.7's clause (d) on automated validation and investigation of unapproved files or unauthorized amendments, combined with regular scanning (f) and defence-in-depth placement (g), can surface some hidden artifacts (e.g., unauthorized hidden files, ADS abuse, hidden accounts), but the technique is explicitly designed to evade the very detection mechanisms A.8.7 mandates, and the clause's primary instrument — signature-based malware detection — targets malware rather than the hiding behaviour itself, leaving a substantial remainder.
- T1564partialresponds — Malware repair software quarantines and removes malicious files, eradicating hidden artifacts when they are recognized as malware, but leaving non-malware hidden artifacts (such as hidden user accounts or modified email rules) unaddressed.
- T1564.001partialdetects — A.8.7's malware detection software (clause f) scans files regardless of hidden attribute and would surface malicious payloads in hidden files, and clause (d) calls for investigating unapproved files, but the specific act of setting hidden attributes is not directly monitored and the technique's purpose is to evade exactly these detection mechanisms, so detection is partial rather than complete.
- T1564.001partialresponds — Malware repair software will quarantine or delete hidden files if they are identified as malicious, eradicating the artifact, but will leave hidden non-malware files (such as staged data or legitimate administrative tools) untouched.
- T1564.003partialdetects — A.8.7's malware detection software can detect some instances of hidden window usage through behavioral analysis (e.g., PowerShell -WindowStyle Hidden command-line flags, CreateDesktop() API calls, registry modifications for window hiding), but the clause's primary focus is on detecting the underlying malware rather than the concealment technique itself, and not all implementations will instrument these specific behaviors.
- T1564.004partialdetects — A.8.7's primary mechanism (malware detection software) is precisely what T1564.004 is designed to evade — the technique's own description names anti-virus and static indicator scanning as defenses it bypasses — but the clause's broader scope (allowlisting per 8.7a, automated validation of data content per 8.7d, defense in depth per 8.7g) catches a slice of ADS abuse, so detection is real but bounded by what the technique is built to hide from.
- T1564.005partialdetects — A.8.7's primary mechanisms (file scanning, application allowlisting, email/web scanning) do not see hidden file systems because the technique is specifically designed to exist outside the standard file system view, but the clause's defence-in-depth principles (g.1) and explicit consideration of attacker evasive techniques (g.2) provide some detection capability through disk-level anomalies, boot-sector/MFT integrity checks, and behavioural monitoring — a slice, not the bulk.
- T1564.005partialprevents — A.8.7's allowlisting (a) and vulnerability reduction (c) can prevent some implementations that require dropping a portable partition image or installing a custom driver/tool, but the technique can also be built using native OS file-system structures (NTFS reserved space, VBR/MFT manipulation, NTFS features) that allowlisting cannot remove, and the technique is explicitly designed to evade the detection-and-repair software (f) that is A.8.7's primary instrument.
- T1564.005partialresponds — Malware repair software can quarantine or delete malicious files that implement a hidden file system, though implementations residing in reserved or unallocated disk space may evade standard eradication.
- T1564.006partialdetects — A.8.7's host-based malware detection cannot see activity inside the virtual instance — which is the technique's explicit purpose — but the control's allowlisting (a), automated validation of software content (d), and network-gateway scanning (g) can surface unauthorized VM software installation, dropped VM artifacts, and anomalous traffic from bridged adapters, so a slice is detectable while the in-guest execution is not.
- T1564.006partialprevents — A.8.7's application allowlisting guidance (a) genuinely prevents unauthorized VM software (VirtualBox, QEMU, dropped binaries) from executing, but native virtualization features like Hyper-V and Windows Sandbox are built into the OS and harder to block, and ESXi-level rogue VMs operate below where A.8.7's host-based controls typically reach, leaving a substantial remainder.
- T1564.007partialdetects — A.8.7's malware detection software is the primary mechanism, and T1564.007 is specifically designed to bypass source-code scanning by hiding payloads in compiled p-code; however, A.8.7's broader scope (defense in depth, consideration of evasive techniques, behavioral detection at gateways and endpoints, and validation of data content) provides partial coverage through behavioral and structural analysis that can surface stomped documents.
- T1564.007partialprevents — A.8.7's primary mechanism is malware detection and repair software scanning VBA source code, which is precisely what VBA Stomping is designed to evade by hiding malicious code in compiled p-code; however, the clause's secondary mechanisms — application allowlisting (a), defense in depth (g), and reducing vulnerabilities exploitable by macros (c) — can prevent the stomped document's macro from executing at all, so a slice of the technique is prevented but the main detection path is bypassed by design.
- T1564.007partialresponds — Malware repair software (such as AV/EDR) can quarantine or delete the malicious document, eradicating the stomped VBA artifact, though the technique's evasive nature may limit how often this response is triggered.
- T1564.009partialdetects — A.8.7's malware detection and repair software (f) and automated validation of software/data content (d) can surface malicious payloads stored in resource forks where the implementation inspects extended attributes and file metadata, but the technique is explicitly designed to evade such scanning, and an A.8.7-conformant implementation that does not specifically inspect resource forks or extended attributes would miss it entirely — the remainder is the slice of implementations whose detection scope reaches this artifact.
- T1564.009partialprevents — A.8.7's application allowlisting (a) and malware scanning (f) can prevent execution of code hidden in resource forks, but the technique is specifically designed to evade detection by abusing a legitimate macOS file system feature, and standard malware detection does not necessarily scan resource forks by default, leaving a meaningful slice where the hiding succeeds.
- T1564.009partialresponds — A.8.7 mandates malware detection and repair software which can quarantine or delete files containing malicious resource forks, effectively eradicating the hidden payload once discovered.
- T1564.010partialdetects — A.8.7's malware detection and repair software (including modern EDR) can detect some instances of PEB manipulation through behavioral analysis, memory scanning, and correlation with related events like process hollowing, but the technique is explicitly designed to evade the process-creation monitoring these tools perform, and many conformant implementations will not catch it.
- T1564.010partialprevents — A.8.7's application allowlisting (a) and defense-in-depth placement (g) can prevent the underlying malware from executing, which forecloses the technique, but T1564.010 is specifically designed to evade the argument-monitoring that A.8.7's detection-and-repair software (f) relies on, and the clause does not address PEB/process-memory manipulation directly, so a substantial slice of instances bypass prevention.
- T1564.011partialresponds — A.8.7's malware repair function forcefully terminates malicious processes and quarantines their files, eradicating the execution even if it ignores standard interrupts, though this response only applies where the execution is identified as malware.
- T1564.014partialdetects — A.8.7 mandates malware detection and repair software that scans files and storage media, but extended attributes are metadata not visible to standard tools and are explicitly excluded from integrity checks that only inspect file content, so detection depends on whether the deployed tooling inspects xattrs, which the clause does not require.
- T1565partialdetects — A.8.7's malware detection and repair software (clause f) and the data-content validation in clause (d) — investigating unapproved files and unauthorized amendments — can surface some instances of data manipulation, particularly those delivered via malware or that alter files on disk; however, T1565 is broader than malware (it includes manipulation through legitimate access, business-logic abuse, and database tampering that leaves no malicious artifact), so A.8.7 catches a slice rather than the technique.
- T1565partialprevents — A.8.7 prevents the subset of T1565 that is delivered via malware (e.g., a troan that tampers with stored data, malicious scripts that alter records), but T1565 is broader than malware: data manipulation can be performed through legitimate credentials, business-logic abuse, SQL injection, or insider access without any malware at all, and A.8.7's scope is explicitly malware protection, leaving those vectors untouched.
- T1565.001partialdetects — A.8.7(d) explicitly requires 'regular automated validation of the software and data content of systems' and 'investigating the presence of any unapproved files or unauthorized amendments,' which is a direct detection mechanism for stored data manipulation; however, this is one bullet among many in a clause primarily scoped to malware, and detection depends on implementation choices about which data stores are validated and whether sophisticated manipulations leave detectable artifacts.
- T1565.001partialprevents — A.8.7's allowlisting (a) and malware detection (f) prevent a slice of stored data manipulation — the instances executed via unauthorized tools or malware — but the technique is broader than malware and can be carried out with legitimate software, manual access, or authorized credentials, which the clause does not address.
- T1565.003partialdetects — A.8.7's malware detection software (item f) and automated validation of software/data content (item d) can surface unauthorized file amendments and known malware used to alter application binaries, but runtime data manipulation that operates purely in memory or modifies legitimate binaries without leaving a detectable artifact falls outside the clause's detection scope.
- T1565.003partialprevents — A.8.7's allowlisting (a) and software/data content validation (d) can prevent some runtime data manipulation vectors—particularly those involving unauthorized binaries or unapproved file changes—but the technique can also be executed by modifying already-allowlisted legitimate applications, changing default file associations, or manipulating data through authorized processes, none of which A.8.7 directly prevents.
- T1565.003partialresponds — Malware repair software can quarantine malicious binaries and kill unauthorized processes performing the runtime data manipulation, but it does not contain or eradicate manipulations executed through legitimate system tools or manual configuration changes.
- T1566partialdetects — A.8.7 explicitly requires malware detection software, regular scans of email/IM attachments/downloads/webpages, collection of new malware intel, and user awareness training to identify suspicious emails/files, all of which surface phishing-borne malware; this is a genuine but minority slice of T1566 because many delivery vectors (e.g. spoofed non-malware links, social engineering without attachments, phone-based lures, or pre-compromise domain acquisition) have no malware artifact for the control to observe.
- T1566partialprevents — A.8.7's email attachment scanning (f.2), webpage scanning (f.3), application allowlisting (a), and user training (m) all act against the malware-delivery aspect of phishing, but the control does not address credential-harvesting phishing, voice phishing, or social engineering without a malicious payload, so the technique is only partially prevented.
- T1566.002partialdetects — A.8.7 explicitly requires scanning email/IM attachments and downloads, scanning webpages when accessed, monitoring for anomalous behaviour via awareness/training/reporting, and collecting/validating malware intelligence, all of which surface spearphishing links (and their payloads) before or during delivery on covered vectors.
- T1566.002partialprevents — A.8.7's malware detection and repair software scans email attachments and downloads before use, and its awareness training helps users identify malicious links, but the technique's reliance on user execution and social engineering means a trained user may still click, so prevention is a reduced rate rather than a closure.
- T1566.003partialdetects — A.8.7's endpoint malware detection and web scanning (f)(1),(3) can surface the malicious payload when it reaches the work computer or a malicious link is clicked, but the technique specifically traverses third-party services (personal webmail, social media) chosen to bypass enterprise email and gateway controls, so the spearphishing messages themselves are not visible to A.8.7's detection mechanisms.
- T1566.003partialprevents — A.8.7's web scanning (f.3), blocklisting of malicious sites (b), application allowlisting (a), and user awareness training (m) all act against the payload and delivery of spearphishing-via-service, but the technique is specifically designed to bypass enterprise email controls by routing through third-party services (personal webmail, social media), so the email-scanning clauses (f.1, f.2) do not reach it and the remaining mechanisms leave a named slice of the technique still possible.
- T1566.004partialprevents — A.8.7's allowlisting (a), blocklisting (b), and awareness training (m) partially constrain downstream actions a voice-phished user might take (installing remote tools, visiting malicious URLs), but the control is framed around malware delivery via email/files/programs and its scanning/detection mechanisms do not reach the voice social engineering itself, which is the technique's core mechanism.
- T1568partialdetects — A.8.7's malware detection software and network gateway scanning (g)1 can surface some dynamic resolution activity — known malware families that beacon to algorithmically-generated domains, malicious C2 webpages, and DNS patterns flagged by anti-malware tools — but the technique's core mechanism (algorithmic generation of C2 addresses, fast-flux resolution, encrypted fallback channels) is a network behaviour that file-centric malware detection and application allowlisting do not directly observe, and the clause's scope is set by risk-based placement (g) rather than mandated depth of network telemetry.
- T1568partialprevents — A.8.7 prevents T1568 indirectly by blocking the malware that would use dynamic resolution (via allowlisting, scanning, and blocklisting), but cannot prevent all instances: dynamic resolution is specifically designed to evade blocklists by algorithmically generating new C2 endpoints, and malware using legitimate software or bypassing allowlisting still reaches the technique.
- T1568partialresponds — A.8.7's malware repair software can quarantine the malicious process performing the dynamic resolution, eradicating the local foothold, though it does not contain the external C2 infrastructure.
- T1568.002partialdetects — A.8.7's malware detection software can identify the malware binary that employs DGAs and block known malicious domains via blocklists (b, f, g), and endpoint scanning can surface processes generating anomalous DNS queries, but the core DGA technique — algorithmic generation of C2 domains — is primarily detected through DNS traffic pattern analysis, which sits more naturally under A.8.16 monitoring activities than under A.8.7's malware-file-scanning scope.
- T1568.002partialprevents — A.8.7's allowlisting, vulnerability management, and email/web scanning can prevent the underlying malware from being installed, which prevents the DGA technique from running; but DGAs are specifically designed to evade the blocklisting (b) and detection (f) controls that are A.8.7's primary mechanisms, and the technique is a C2 behavior that operates after compromise, so a substantial slice of DGA usage survives the control's preventive surface.
- T1569partialprevents — A.8.7's application allowlisting (a) can prevent the creation and execution of unauthorized services, and malware detection (f) can block malicious payloads, but T1569 also encompasses abusing existing legitimate system services (e.g., via sc.exe, PsExec) which allowlisting permits by design and detection cannot reliably preempt.
- T1569partialresponds — Malware repair software can quarantine and remove malicious executables launched via system services, eradicating the payload, but this does not contain the abuse of services to run legitimate or living-off-the-land commands.
- T1569.001partialdetects — A.8.7's guidance includes detection mechanisms — application allowlisting (a), automated validation of software and data content investigating unapproved files (d), and malware detection and repair software (f) — that can surface launchctl abuse by flagging unauthorized LaunchAgent/LaunchDaemon plist creation or malicious payloads being loaded; however, the clause is primarily preventive, explicitly states detection software alone is not usually adequate, and detection of abuse of a legitimate macOS utility requires behavioral monitoring whose depth is set by implementation rather than mandated by the clause.
- T1569.001partialresponds — Malware protection and repair software can terminate malicious processes spawned via launchctl and quarantine the associated payloads, though it may not remediate interactive abuse of the native tool itself.
- T1569.002partialdetects — A.8.7's malware detection software and application allowlisting can surface malicious payloads and unauthorized services created via the service control manager, but T1569.002's core mechanism is abuse of legitimate Windows utilities (sc.exe, services.exe, PsExec) to execute commands — the technique's signature is the use of built-in binaries, which malware protection is not designed to detect, so only a slice of instances is caught.
- T1569.002partialprevents — A.8.7's allowlisting (a) and malware detection (f) can constrain unauthorized service creation and catch malicious payloads executed via services, but the service control manager is a required Windows component that cannot be removed, so the underlying mechanism remains abusable.
- T1569.002partialresponds — A.8.7's malware repair software can quarantine malicious payloads and terminate their processes once executed via the service control manager, but this automated response does not extend to the execution of non-malware commands or native system utilities.
- T1569.003partialdetects — A.8.7's malware detection software and unauthorized-software controls can surface malicious systemd service files or payloads executed via systemctl, but the core technique abuses a legitimate system binary that AV is not designed to flag, so detection covers artifacts rather than the abuse pattern itself.
- T1570partialdetects — A.8.7's malware detection software scans files received over networks and could detect known malicious payloads being transferred laterally, but the technique's core mechanism—moving tools between already-compromised systems via legitimate protocols (SMB, RDP, scp, rsync, etc.)—is not what the clause targets; legitimate dual-use tools and custom malware evade content-based detection, and the clause's scope is malware content rather than lateral-movement patterns.
- T1570partialprevents — A.8.7's malware detection at network gateways and endpoints (clauses f, g) can scan files in transit and block malicious payloads, preventing some lateral tool transfers where the transferred content is recognized as malware; however, T1570 uses legitimate file-sharing protocols (SMB, RDP, scp, rsync, curl, etc.) and A.8.7 does not restrict those mechanisms, and transfers of legitimate tools repurposed by the adversary are not caught, so only a slice of instances is prevented.
- T1570partialresponds — Malware repair software can quarantine or delete malicious files after they have been transferred, eradicating the artifact of the transfer, but this leaves dual-use or native administrative tools untouched.
- T1571partialdetects — A.8.7's malware detection at network gateways (g.1) scans traffic across application protocols and can detect malware payloads transmitted over non-standard ports if configured for defence in depth, but the clause's primary mechanism is malware-content detection rather than network-behavior analysis, so the non-standard port usage as a bypass technique is only incidentally surfaced.
- T1572partialdetects — A.8.7's malware detection and repair software at endpoints and gateways can flag tunneling tools (SSH clients, DoH resolvers) as potentially unwanted programs and may surface anomalous encapsulated traffic during content scanning, but the clause is built around file/email/web content scanning rather than network protocol analysis, and protocol tunneling is specifically designed to evade such detection by blending with legitimate traffic or using encryption.
- T1572partialprevents — A.8.7's malware detection and application allowlisting can prevent tunneling that requires malicious or unauthorized software on the endpoint, but the technique commonly abuses legitimate built-in tools (SSH clients, DNS resolvers) and standard protocols, which allowlisting cannot remove and which malware detection does not flag as malicious — so a meaningful slice is prevented but the technique as a whole is not.
- T1573partialprevents — A.8.7 prevents T1573 only indirectly by blocking the underlying malware from being installed (via allowlisting, blocklisting, scanning, vulnerability reduction); the technique itself concerns encryption of the C2 channel, which A.8.7 does not address, and the guidance explicitly acknowledges that attackers use encryption to evade malware detection and that detection software alone is not usually adequate.
- T1573partialresponds — A.8.7's repair function quarantines or removes the underlying malware, which eradicates the encrypted channel it generates, though it does not directly block the network traffic or address non-malware implementations.
- T1574partialprevents — A.8.7's application allowlisting (8.7a) blocks some T1574 variants where the malicious payload is a separate unauthorized executable, and defense-in-depth placement (8.7g) plus vulnerability reduction (8.7c) constrain others; however, T1574 is explicitly designed to evade application control by hijacking the execution flow of already-allowed programs (DLL search-order hijacking, path interception, services registry permissions, etc.), so the technique's core mechanism survives the control's primary defense and the remainder is nameable rather than open-ended.
- T1574partialresponds — Malware protection and repair software can quarantine or remove the malicious payloads (such as dropped DLLs or executables) used to hijack execution flow, eradicating the foothold, though it typically does not revert the configuration changes that enabled the hijack.
- T1574.001partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent some DLL hijacking by blocking unauthorized DLLs and scanning for malicious payloads, but the technique is specifically designed to evade these defenses — the legitimate loader program is trusted, malicious DLLs are often encrypted/packed until loaded into the trusted process, and standard allowlisting typically governs executables rather than DLL-level loading, leaving a named remainder.
- T1574.001partialresponds — Malware repair software can quarantine the malicious DLL and terminate the hijacked process once identified, eradicating the immediate artifact but not remediating the vulnerable application or search order.
- T1574.004partialdetects — A.8.7's detection mechanisms (AV/EDR scanning, application allowlisting per 8.7a, regular automated validation of software content per 8.7d investigating unapproved files) can surface some dylib hijacking instances, but the technique is explicitly designed to evade security products by inheriting a legitimate process's execution context, and the clause's guidance is general rather than tailored to macOS dylib search-path abuse.
- T1574.004partialprevents — A.8.7's allowlisting (a) and vulnerability reduction (c) provide some preventive coverage against unauthorized dylibs being placed in search paths, but the technique specifically exploits legitimate application search behavior and is documented as evading security products, and the clause does not address macOS-specific mechanisms (code signing, Gatekeeper, notarization) that would be most effective against dylib hijacking.
- T1574.005partialdetects — A.8.7's malware detection software (f) and automated validation of software content (d) can detect the malicious binary when placed in the installer directory or when executed, but the technique is specifically designed to evade detection by exploiting installer-trusted locations and file permission weaknesses that A.8.7 does not directly address; detection depends on AV signatures and behavioral detection catching the payload, leaving a bounded remainder where novel or trusted-looking binaries are not flagged.
- T1574.005partialprevents — A.8.7's allowlisting (a), software-content validation (d), and malware detection (f) can prevent or catch some instances of replaced installer binaries, but the clause does not directly address the file-system permission weakness on installer directories (typically %TEMP% subdirectories) that is the root enabler of T1574.005; that hardening lives in A.8.1/A.8.18/A.8.32, leaving A.8.7 with a slice of coverage rather than the bulk.
- T1574.005partialresponds — Malware repair software can quarantine the malicious binary dropped into the installer directory and kill the hijacked process, but this containment and eradication only occurs if the payload is recognized as malware.
- T1574.006partialdetects — A.8.7's malware detection and repair software (clause f) and defense-in-depth placement (clause g) can surface many dynamic linker hijacking attempts via behavioral rules on LD_PRELOAD/DYLD_INSERT_LIBRARIES, unexpected library loads, and file scans of the planted shared object; however, the technique is specifically engineered to evade these controls by hooking system functions (execve, readdir, ldd) to scrub its own artifacts, and in-memory or encrypted payloads defeat signature-based detection, leaving a bounded remainder that the control does not see.
- T1574.006partialprevents — A.8.7's application allowlisting (a) and malware detection (f) provide meaningful prevention against the malicious shared library being written or executed, but the dynamic linker mechanism is fundamental to Linux/macOS and cannot be removed, the technique is specifically designed to evade detection by hooking system functions like execve and readdir, and the adversary must already have access to set environment variables — a precondition A.8.7 does not address.
- T1574.006partialresponds — Malware repair software can quarantine or remove the malicious shared library, but may not revert the environment variable modifications (such as LD_PRELOAD) that constitute the hijacking mechanism itself.
- T1574.007partialdetects — A.8.7's guidance includes application allowlisting (item a) and automated validation of software content investigating unapproved files (item d), both of which can detect unauthorized binaries placed in PATH directories. However, the technique also involves modifying the PATH environment variable itself, which A.8.7's malware-focused controls do not directly address, and novel or renamed legitimate-looking payloads can evade signature-based detection (item f), leaving a bounded remainder.
- T1574.007partialprevents — A.8.7's allowlisting (a) and malware detection (f) can prevent the malicious binary from executing or detect it, but the technique is specifically designed to evade name-based controls by using legitimate binary names, and the clause does not mandate hash/path-based allowlisting or address PATH variable integrity directly.
- T1574.007partialresponds — Malware repair software can quarantine and remove the malicious binary placed in the intercepted path, eradicating the payload, though it does not revert modifications made to the PATH variable itself.
- T1574.008partialdetects — A.8.7 mandates malware detection and repair software that scans files on access and during regular automated validation, which can surface a malicious executable placed via search order hijacking when it is written to disk or executed, but the control's detection scope is file-based and does not guarantee visibility into the process creation chain or the search order logic that makes the hijacking possible.
- T1574.008partialprevents — A.8.7's allowlisting (a), file/content validation (d), and malware detection (f) all bear on the technique — the malicious file is an unauthorized executable placed in a directory — but the technique exploits a specific Windows search-order behavior that A.8.7 does not directly target, allowlisting has documented exceptions (i), and the payload can be named to masquerade as a legitimate system utility, leaving a bounded but real remainder.
- T1574.008partialresponds — Malware repair software can quarantine or delete the malicious executable placed to hijack the search order if it is recognized as malware, eradicating the specific payload but not addressing the vulnerable calling program or the permissions that allowed the file write.
- T1574.009partialdetects — A.8.7's malware detection and repair software (item f) and investigation of unapproved files (item d) can surface the adversary-placed executable that path interception causes to run, but the control does not look for the unquoted-path vulnerability itself or the path-resolution hijack as a technique — detection is incidental to payload recognition and does not cover the underlying misconfiguration.
- T1574.009partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can block the adversary's planted executable from running, but the technique exploits a specific configuration weakness (unquoted service/shortcut paths) that A.8.7 does not directly address — fixing unquoted paths belongs to configuration/vulnerability management (A.8.8/A.8.9) — so prevention is indirect and depends on which sub-mechanisms are implemented.
- T1574.009partialresponds — Malware repair software can quarantine or delete the malicious executable placed in the intercepted path if it is recognized as malware, eradicating the artifact but leaving the underlying unquoted path vulnerability intact.
- T1574.010partialdetects — A.8.7's malware detection software (clause f) and regular automated validation of software content (clause d) can surface the replaced binary or unauthorized file amendments, but the technique specifically exploits weak file permissions on service binaries — a configuration weakness A.8.7 does not directly monitor — so detection depends on signature coverage and integrity checks the clause does not mandate.
- T1574.010partialprevents — A.8.7's malware detection software (f), application allowlisting (a), and automated validation of software content (d) provide some prevention against malicious binaries being placed or executed, but the specific weakness exploited by T1574.010 — improper file system permissions on service binaries or their directories — is not addressed by the clause's guidance, and allowlisting by service path will not catch a replaced legitimate binary.
- T1574.011partialdetects — A.8.7's malware detection and repair software (clause f) would surface the malicious DLL or executable when the hijacked service loads it, and clause (d)'s automated validation of software and data content can flag unauthorized amendments including registry changes, but the technique's core mechanism — abusing weak ACLs on service registry keys to redirect execution — is a permissions/configuration abuse that A.8.7's malware-focused detection is not designed to catch directly, leaving the registry hijack itself largely unobserved until the payload runs.
- T1574.011partialprevents — A.8.7's application allowlisting guidance (item a) can prevent the malicious payload from executing even when the service registry is hijacked, but the clause does not address the underlying weak permissions on HKLM\SYSTEM\CurrentControlSet\Services that enable the hijack itself, and allowlisting is bypassable in practice.
- T1574.011partialresponds — Malware repair software can quarantine the malicious executable or DLL referenced by the hijacked registry key, eradicating the payload but leaving the underlying permissions weakness unresolved.
- T1574.012partialdetects — A.8.7's malware detection and repair software (AV/EDR) can surface the malicious profiling DLL on disk and behavioral rules can flag COR_PROFILER environment variable abuse, but the technique deliberately leverages a legitimate .NET framework feature and the DLL may be unknown or appear benign, so detection is a slice rather than a bounded remainder of a working mechanism.
- T1574.012partialprevents — A.8.7's application allowlisting (8.7a) and malware detection (8.7f) can prevent the malicious profiling DLL from loading or executing, but COR_PROFILER specifically abuses the trusted .NET CLR to load DLLs into every .NET process, and the technique supports in-memory registration that bypasses registry-based controls, so a meaningful slice of the technique survives the control's mechanisms.
- T1574.012partialresponds — A.8.7's malware detection and repair software (f) and procedures for recovering from malware attacks (l, j) provide a response capability against the malicious DLL payload, but the technique's persistence mechanism (COR_PROFILER environment variable) and its ability to hook and impair .NET defenses are not fully addressed, leaving the eradication incomplete.
- T1574.013partialdetects — A.8.7's malware detection software (broadly construed to include modern EDR with behavioral analysis) can surface anomalous memory operations like WriteProcessMemory into another process's PEB or unusual NtQueryInformationProcess calls, but the technique is explicitly designed to evade detection by masking execution under a legitimate process, and A.8.7's primary mechanism — file/email/web scanning — does not apply to this in-memory callback-table hijack, so what is caught is a slice rather than the bulk.
- T1574.013partialprevents — A.8.7's application allowlisting (a) and malware detection (f) can prevent the initial payload that would perform the hijack from executing, but the technique's core mechanism — in-memory manipulation of a running process's PEB via WriteProcessMemory — operates after code is already running and is explicitly designed to evade detection from security products, which A.8.7's primary instrument cannot reach.
- T1574.013partialresponds — Malware protection tools can contain the threat by terminating the compromised process, but this is only partial because the technique explicitly masks execution within a legitimate process to evade such security products.
- T1574.014partialdetects — A.8.7's malware detection software (clause f) and application allowlisting (clause a) can surface some instances of AppDomainManager injection — known malicious assemblies, unauthorized .config modifications, or unapproved files flagged by automated validation (clause d) — but the technique is specifically designed to evade traditional malware detection by loading attacker-controlled assemblies into legitimate .NET processes, and A.8.7's scope is general malware protection rather than specialised .NET runtime or process-injection telemetry.
- T1574.014partialprevents — A.8.7's application allowlisting (a), configuration validation (d), and malware detection (f) all bear on AppDomainManager hijacking, but the technique abuses legitimate .NET framework functionality that is typically required for business operations and cannot be removed, so what is prevented is a slice rather than the bulk with a bounded remainder.
- T1574.014partialresponds — Malware repair software can terminate the compromised process and quarantine the malicious assembly used in the injection, providing containment and eradication of the threat.
- T1598.002partialdetects — A.8.7's email attachment scanning (f.2) detects malicious attachments used in spearphishing, but T1598.002's primary goal is information elicitation through social engineering rather than malware execution, and encrypted attachments, HTML smuggling, and novel payloads can evade signature-based detection.
- T1598.002partialprevents — A.8.7's email attachment scanning (clause f.2) and malware-focused awareness training (clause m) prevent variants where the attachment carries executable malware, but T1598.002 is a reconnaissance technique whose goal is information elicitation, not malware execution; variants using fake login portals (HTML smuggling) or fillable forms without malware bypass malware scanning, and the adversary's probing still occurs even when individual lures fail.
- T1598.003partialdetects — A.8.7's email and web scanning (f.2, f.3) plus blocklisting of known malicious websites (b) can detect spearphishing links that lead to malware payloads or known-malicious destinations, but T1598.003 is fundamentally about social engineering to elicit credentials or other information — the core pretext, credential-harvesting forms without malware, and QR-code/'browser-in-the-browser' lures fall outside A.8.7's malware-focused detection scope.
- T1598.003partialprevents — A.8.7's blocklisting of malicious websites (b) and user training to identify malicious emails (m) partially prevent T1598.003 from succeeding — the reconnaissance email may still arrive but the credential-harvesting site can be blocked and trained users may not click — though the clause is scoped to malware rather than phishing-for-information, and credential-harvesting links often evade malware-focused detection.
- T1601partialdetects — A.8.7 mandates regular automated validation of software and data content (d) and malware scanning of files before use (f), which can surface an unauthorized modification to a monolithic system image, but the control's detection scope is bounded by the placement and configuration of tools (g) and the evasive techniques of attackers (g2), so what is caught is a slice the implementer chose rather than a bounded remainder of a mechanism that mostly works.
- T1601.001partialdetects — A.8.7's primary mechanism (malware detection and repair software) does not run on network devices, so the bulk of the clause cannot observe T1601.001; however, item (d) on automated validation of software and data content for unauthorized amendments and item (g) on defense-in-depth at network gateways provide partial detection capability for anomalous image changes and related traffic, leaving a bounded remainder rather than full coverage.
- T1610partialdetects — A.8.7's malware detection software can catch malicious payloads within container images or at runtime, and application allowlisting can flag unauthorized software, but the deployment action itself (creating/starting containers via orchestration APIs) is largely outside the clause's detection scope, which targets malware content rather than container orchestration events.
- T1612partialdetects — A.8.7's malware detection software (item f) and automated validation of software content (item d) can surface the malicious payload downloaded from C2 during the build or the resulting image when scanned, but T1612 is explicitly designed to evade defenses that monitor image retrieval, and the on-host build activity itself is not the primary target of A.8.7's file/email/web scanning mechanisms.
- T1620partialdetects — A.8.7's detection mechanisms (f) are primarily file-based scanning — data, email attachments, downloads, webpages — and T1620 is specifically designed to evade these by loading code directly into process memory without creating on-disk artifacts; the clause's awareness of evasive techniques (g.2) and modern EDR-style behavioral detection provide some coverage, but the technique's fileless, in-memory nature means most of A.8.7's prescribed scanning does not reach it.
- T1620partialprevents — A.8.7's allowlisting (8.7a) and scanning of inbound files/email/web (8.7f) can block the initial dropper or unauthorized interpreter that delivers the reflective payload, but the technique is explicitly designed to evade file-based and process-based detection by loading code directly into a running process's memory, which sits outside the clause's primary mechanisms.
- T1620partialresponds — A.8.7 explicitly requires procedures, training, awareness, and business continuity plans to deal with and recover from malware attacks once underway (l, j, m), which engages the containment/eradication/response slice of reflective code loading when it manifests as malware; this is only partial because the control's dominant mechanisms are preventive/detective (scanning, allowlisting, vulnerability reduction) and many fileless in-memory cases produce no observable artifacts for the response procedures to act on.
- T1622partialdetects — A.8.7's malware detection and repair software (especially EDR components) can detect some anti-debugging API calls and behaviors (e.g., IsDebuggerPresent, NtQueryInformationProcess, OutputDebugStringW flooding), but the technique is specifically designed to evade analysis, and A.8.7's primary purpose is detecting malware payloads rather than the specific defensive-evasion behavior; sandbox/analysis environments (outside A.8.7's scope) are the more natural vantage point for observing this technique.
- T1622partialresponds — Malware detection and repair software can quarantine the malicious process, eradicating the artifact performing the debugger evasion.
- T1647partialdetects — A.8.7 mandates malware detection and repair software, application allowlisting, and automated validation of software/data content (clauses a, d, f), and modern endpoint protection does include behavioral monitoring that can surface suspicious plist writes — but plist modification is a configuration change rather than a malware payload, the technique is explicitly described as evading system defenses, and A.8.7 sets a framework rather than mandating plist-specific integrity monitoring, so detection is a slice of instances rather than the bulk.
- T1647partialprevents — A.8.7's application allowlisting (a), malware detection (f), and validation of software/data content (d) can prevent some plist-based persistence and unauthorized application execution, but plist modification typically targets legitimate system files and existing applications using native macOS mechanisms, so the technique's core vectors—modifying LSUIElement, LSEnvironment, or Launch Agent/Daemon plists on an already-accessible system—remain reachable and the prevention is a slice rather than the bulk.
- T1659partialdetects — A.8.7's malware detection at network gateways (g.1) and webpage scanning (f.3) can catch injected payloads carrying known malicious signatures, but content injection specifically operates at upstream channels (ISP, compromised intermediaries) outside the organization's gateway vantage point, and the injected content is designed to appear legitimate, so the injection mechanism itself is not what endpoint or gateway AV detects — only a slice of payloads is caught.
- T1659partialprevents — A.8.7's application allowlisting (a) and webpage/email scanning (f) can prevent some injected content from executing or being delivered, but content injection fundamentally occurs at upstream network layers (ISP-level, in-transit manipulation) that endpoint malware controls cannot reach, leaving the core attack vector outside the control's scope.
- T1674partialdetects — A.8.7's malware detection software (clause f) can surface known input-injection malware such as BackSwap-style tools that monitor the Windows message loop and simulate keystrokes, but physical HID attacks (malicious USB devices emulating keystrokes) and sophisticated/fileless variants evade signature-based detection, and the clause's scope is malware rather than the keystroke-simulation behaviour itself.
- T1674partialprevents — A.8.7's layered malware defenses (allowlisting, email/attachment scanning, web scanning, detection tools) prevent the software-based delivery and execution of keystroke-injection malware, but the clause does not address physical HID devices (malicious USB emulators) which bypass software controls entirely, and once keystrokes are injected the technique can leverage legitimate system tools like PowerShell that allowlisting cannot remove.
- T1674partialresponds — A.8.7's malware repair capabilities can quarantine and eradicate the malicious software tooling used to simulate keystrokes, but this does not contain or eradicate input injection performed via physical Human Interface Devices (HIDs).
- T1677partialdetects — A.8.7's mechanisms have partial observability of T1677: malware detection software (item f) can flag injected payloads containing known signatures, application allowlisting (item a) can detect unauthorized tools running in the build environment, and automated validation of software/data content (item d) can surface unauthorized amendments to CI configuration files. However, the core of T1677 — manipulating CI/CD configs, scripts, and pull-request inputs to execute within the pipeline itself — operates in a build-environment context that A.8.7's traditional endpoint and gateway scanning does not specifically cover, and novel or fileless payloads (e.g., credential-exfiltration commands) evade signature-based detection.
- T1678partialresponds — Malware repair and isolation capabilities can quarantine the malicious process performing the delay, containing and eradicating it once underway, though the technique's evasive nature means this response is not guaranteed.
- T1684partialdetects — A.8.7's malware detection and repair software scans files, emails, and webpages for malware before use, which can detect malicious payloads delivered through social engineering, but the technique's core manipulation of human trust and the social engineering interaction itself occur outside the software's scanning scope.
- T1684partialprevents — A.8.7's training clause (m) addresses the malware-delivery slice of T1684 (infected emails, files, programs), and its technical controls (allowlisting, blocklisting, malware detection) prevent execution even when a user is socially engineered into installing a payload; but T1684's broader outcomes — password resets, MFA changes, financial approvals, disclosure of sensitive information — sit outside A.8.7's scope, so only a slice is prevented.
- T1685.003partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, but T1685.003 manipulates the tool's own UI to falsely show a healthy status, an in-memory deception that file-based and network scanning do not observe; detection is limited to the slice where behavioral or integrity monitoring is implemented under the control's broader guidance.
- T1685.004partialdetects — A.8.7 mandates malware detection and repair software that scans files, email, and webpages, which can surface the malware that performs the audit system tampering, but the control's scanning scope does not inherently monitor the integrity of the Linux Audit system configuration or daemon itself, so detection of the disabling action is incidental rather than systematic.
- T1687partialdetects — A.8.7's layered architecture (defense in depth per g.1), software/data validation (d), and monitoring of security tool integrity can surface exploitation of defensive components, but the primary detection mechanism — the malware detection and repair software itself — is the target of T1687, and the clause is framed around protecting against malware rather than protecting the protection mechanisms, so detection is a bounded slice rather than a complete coverage.
- T1687partialprevents — A.8.7 clause (c) requires reducing vulnerabilities exploitable by malware through technical vulnerability management (cross-referencing 8.8), which can include patching the security tools themselves and thereby prevent exploitation of known flaws in EDR/AV; however, the clause's primary focus is USING malware detection software rather than HARDENING it, zero-day exploitation of defensive tools remains unaddressed, and the protection is incidental rather than direct.
Prevented OWASP Web Top 10 (2025) risks (9)
OWASP Web Top 10 (2025) risk categories this ISO control helps prevent or mitigate — our analysis (mostly llm_verified).
Direction: ← other covers this;
→ this covers other (F/M/P = full / mostly /
partial). gov = governs / implements (a mandate, not coverage).
Why these map — rationale
- A02partialblocks — A.8.7 guidance explicitly includes reducing exploitable vulnerabilities (c), preventing unauthorized software via allowlisting (a), blocking malicious sites (b), scanning for unapproved changes (d), and hardening placement/configuration (g), which directly close many common misconfiguration vectors that malware exploits; however, the control's dominant focus is runtime malware detection/repair and awareness rather than systematic secure configuration of defaults, cloud settings, or frameworks, leaving a substantial remainder of the class untouched.
- A05partialblocks — A.8.7 explicitly prevents malware vectors that deliver injection (e.g. via malicious websites, email attachments, untrusted files, and unpatched vulnerabilities), but leaves the dominant injection shapes (SQLi, command injection, XSS in trusted input paths) untouched.
- A08partialblocks — A.8.7 prevents a slice of A08:2025 via malware rules that block unauthorized software (a), reduce exploitable vulns (c), scan/validate data and software (d,f), and protect external sources (e), but leaves most of the class (insecure deserialization, unsigned updates, CI/CD pipeline tampering) untouched.
Control IDs, short titles and the structured attribute table (control type, CIA properties, cybersecurity-concept, operational capability, security domain) are facts from ISO/IEC 27001:2022 Annex A / ISO/IEC 27002:2022. The full implementation guidance prose lives in ISO/IEC 27002:2022 — not reproduced here. Cross-walks to NIST 800-53, NIST CSF 2.0, OWASP ASVS, CWE, MITRE ATT&CK and OWASP Web Top 10 are our own analysis (mostly llm_verified), not an ISO, NIST, MITRE or OWASP product — how ours compare.