A.8.21 Technological
Security of network services
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 (26)
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
- SC-7partialcovers — A.8.21's broad requirement to ensure security when using network services (including their secure configuration and monitoring) addresses only a slice of SC-7's specific boundary-protection, subnetwork segmentation, and managed-interface mandates; the bulk of SC-7's technical controls sit elsewhere.
- CA-3domain mandateimplements — A.8.21's governance mandate for secure network service usage directly encompasses the domain of controlled, documented information exchanges with external systems that ca-3 operationalizes
Aligned NIST CSF 2.0 outcomes (23)
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
- DE.CM-01impliedimplements — DE.CM-01 names network monitoring as the required outcome; A.8.21's measures for securing network services (including monitoring and logging of usage) give operational effect to that outcome within the network-services domain, but the link is inferential rather than by explicit naming of monitoring.
- PR.AA-03domain mandateimplements — A.8.21 operationalizes network-service security which directly includes authenticating users/services/hardware as a core means within that domain (per its implementation guidance on authentication and access to services), though PR.AA-03 is not named verbatim
- PR.IR-01domain mandateimplements — A.8.21 operationalizes network-service security (including access/use restrictions) which sits inside the PR.IR-01 protection outcome; the link is by shared domain rather than explicit naming of the control's subject.
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).
Why these map — rationale
- V13.2.4mostlycovers — The ISO control's requirement to define and enforce an allowlist of permitted networks and services directly implements the ASVS mandate that applications only communicate with explicitly approved external resources.
- V13.2.5mostlycovers — Specifying which resources a server may contact and restricting outbound connections aligns with the ISO guidance on formulating rules for allowed network access and technological controls.
Related weaknesses / CWE (3255)
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-319mostlyblocks — A.8.21 mandates identifying/implementing network service security features (explicitly including encryption and connection controls), formulating rules that cover authentication/authorization/monitoring/VPNs, and ensuring providers do the same; this stops cleartext transmission from being exploitable in the large majority of network channels, with a bounded remainder (e.g. internal non-networked flows or unaddressed legacy protocols).
- CWE-319mostlyremoves — A.8.21 mandates that security features for network services — including encryption and technical parameters for secured connection — be identified and implemented, so applying it where it was not applied before removes cleartext transmission over those network services, which is the bulk of CWE-319; the bounded remainder is non-network channels (hardware debug paths, JTAG, internal hardware networks) that fall outside the clause's network-services scope.
- CWE-1125partialblocks — A.8.21 mandates authentication, authorization, encryption and connection controls on network services, which blocks exploitation of those specific attack-surface elements, but the weakness (an architecturally excessive attack surface) remains present and the clause does not reach non-network-service entry points or reduce the surface's quantitative measure.
- CWE-1327partialblocks — A.8.21's rules on network service access (allowed networks, authorization procedures, network management controls, monitoring) provide a framework that can include firewalls or segmentation blocking exploitation of a 0.0.0.0 binding, but the clause does not directly mandate binding-address restrictions and the blocking depends on referenced network controls rather than the clause itself.
- CWE-1428partialremoves — A.8.21 mandates encryption as a security feature of network services and technical parameters for secured connections, so applying it where it was not applied before requires HTTPS for network service communications, removing the HTTP reliance in that scope. The remainder is HTTP usage outside network services — internal APIs, inter-component calls, links in generated content — which the clause does not reach.
- CWE-1428partialblocks — A.8.21's guidance on encryption and connection controls for network services can require HTTPS and block HTTP at the network service level, denying exploitation over that path, but the product's HTTP usage remains in code/design and not all network services (third-party, external) are under the clause's control, leaving a nameable remainder.
- CWE-220partialblocks — A.8.21 mandates authentication, authorization and monitoring for network services including FTP, which can block untrusted parties from reaching the FTP root where sensitive data sits; however, the clause governs network service access in general and does not address the placement of sensitive data under the FTP root itself, so exploitation remains possible where authorization is misconfigured, where authenticated users retain read access to the sensitive files, or where the data's presence under FTP root is the design defect the clause does not reach.
- CWE-284partialblocks — A.8.21 directly addresses access control for network services (authorization procedures, authentication requirements, network connection controls, means of access), blocking exploitation of CWE-284 in that slice, but CWE-284 is broader than network services and includes application-level authorization, file permissions, API access control, and intra-application privilege failures that A.8.21 does not reach.
- CWE-288partialblocks — A.8.21's authentication requirements (item b), authorization procedures (item c), and network connection controls (item d) can be implemented to require authentication on all access paths to a network service, which blocks exploitation of an unauthenticated alternate path; however, the underlying architectural flaw in the product remains, and the clause cannot redesign third-party or COTS products where the alternate path is hardcoded into the design, leaving a nameable remainder.
- CWE-294partialblocks — A.8.21 requires security features for network services including authentication and encryption, which can block capture-replay during network transit, but the clause does not specifically mandate replay-resistant authentication protocols and the actual protocol design lives in other controls (A.8.24 cryptography, A.8.28 secure coding); encryption alone does not stop replay of an encrypted message, so the weakness remains exploitable where the authentication protocol lacks nonces/timestamps.
- CWE-306partialremoves — A.8.21 mandates authentication requirements for accessing network services and requires security measures (including authentication) to be identified and implemented, so applying it where it was not applied before results in authentication being added to network services that lacked it, removing those CWE-306 instances; however, A.8.21's scope is limited to network services while CWE-306 also covers non-network critical functions (UART debug ports, local APIs, backdoor channels), and the clause governs the requirement rather than directly implementing authentication, leaving a nameable remainder.
- CWE-308partialblocks — A.8.21 mandates authentication requirements (item b) and authentication technology (security features item a) for network services, so a well-implemented clause blocks exploitation of single-factor auth in the network-services scope; CWE-308 also arises in non-network-service authentication contexts the clause does not reach, leaving a nameable remainder.
- CWE-309partialblocks — Network access rules, VPN requirements, and authorization controls in A.8.21 can keep some unauthorized remote attackers from reaching a password-authenticated service, but the underlying password-only mechanism remains exploitable by anyone with permitted network access or through allowed public services.
- CWE-353partialblocks — A.8.21's guidance on security features of network services (authentication, encryption, network connection controls) and technical parameters for secured connections can require integrity verification as a feature of allowed network services, which blocks exploitation during transit over those services; however, the clause governs service-level security features rather than mandating integrity checks at the protocol layer, so the underlying protocol weakness remains exploitable in contexts outside the governed service path.
- CWE-354partialblocks — A.8.21 mandates security features of network services including encryption and connection controls, and technical parameters for secured connections, which can require integrity-protected transit (e.g., TLS) that blocks exploitation of the missing checksum check during network transmission; however, the application-level checksum validation defect itself remains and can be exploited where data is processed before validation, where the network service is bypassed, or where the checksum serves purposes beyond transit integrity.
- CWE-419partialblocks — A.8.21 governs the security of network services through agreements, rules, and feature selection, but CWE-419 describes an unprotected channel that already exists in the product's design; the control does not interpose a mechanism that stops an attacker from exploiting that channel.
- CWE-419partialremoves — A.8.21 mandates security features (authentication, encryption, connection controls) for network services and requires providers to implement them, so applying it to existing network services with unprotected primary channels removes those CWE-419 instances; however, the clause only reaches channels that are part of network services, not all primary channels (e.g., local admin interfaces), and removal depends on provider compliance rather than direct code change.
- CWE-420partialblocks — A.8.21's requirements for authentication, authorization and monitoring of network services would block exploitation of CWE-420 instances where the alternate channel is itself a network service interface (e.g., the socket/TCP/FTP/named-pipe examples in the CWE), but the clause does not specifically mandate equivalent protection across channels and does not reach non-network alternate channels such as the IBUS/DBUS or shadow-register examples, leaving a nameable remainder.
- CWE-5partialblocks — A.8.21 requires encryption and secured-connection parameters for network services, which can block plaintext transmission at the network service level, but CWE-5 is a J2EE application misconfiguration (the app accepts HTTP) that A.8.21 does not reach into the application to fix; the weakness remains present and exploitable where the network service is bypassed or the encryption requirement is not enforced end-to-end.
- CWE-521partialblocks — A.8.21 item (b) covers authentication requirements for accessing network services, which can include password strength requirements, blocking exploitation of weak passwords on those services; the remainder is bounded and nameable -- passwords for non-network-service contexts (local accounts, application logins, etc.) are outside the clause's scope, mirroring the A.8.20 -> CWE-523 partial pattern where the control covers part of the surface.
- CWE-523partialblocks — A.8.21 mandates network-level encryption (e.g., VPN) and connection controls, which masks credentials during a portion of their transit, but does not require application-layer protection (e.g., HTTPS) for the full path.
- CWE-524partialprevents — A.8.21's caching guidance (item c under security features) lets the organization configure network-service caching parameters so users can choose caching based on confidentiality requirements, which prevents some instances of CWE-524 at the network service layer (e.g., CDN caching of sensitive responses); the remainder is application-level caching decisions — in-memory caches, HTTP response headers, query caches — which the clause does not reach.
- CWE-524partialremoves — A.8.21's caching guidance explicitly addresses network-service caching (e.g., CDN) parameters so users can choose caching behaviour against confidentiality requirements, which removes instances of CWE-524 at the network-service level (the observed CVE-2024-30127 Cache-Control case sits squarely here); it does not reach application-level caching defects where sensitive data is cached in application memory or other first-party stores, which is the broader class.
- CWE-524partialblocks — A.8.21's caching provision (item c) addresses network-service caches (e.g., CDN) and their parameters, allowing the organization to configure caching to avoid storing sensitive content or to restrict access to it, which blocks exploitation via that path; however, CWE-524 also covers application-level caches (in-memory, database, framework caches) and the HTTP Cache-Control header behavior seen in CVE-2024-30127, which A.8.21 does not reach, leaving a nameable remainder of the weakness still exploitable.
- CWE-552partialblocks — A.8.21's network service controls (authentication, authorization, network access restrictions, procedures to restrict access to network services) can block external parties from reaching the service that exposes the files, but the underlying file/directory access control weakness is at the storage layer and is not addressed by this clause.
- CWE-651partialblocks — A.8.21's network access controls, authentication requirements and authorization procedures can block unauthorized network access to the WSDL endpoint, but the weakness persists for any party granted access (the WSDL still contains the sensitive method/parameter information that aids guessing of other resources), and the clause does not address the WSDL's content or whether sensitive method names should be exposed at all.
- CWE-654partialblocks — A.8.21 mandates authentication requirements for network services (item b) and can require additional contextual factors like time and location (item f), so a well-implemented clause blocks single-factor decisions at the network-service access boundary; the remainder is bounded and nameable -- application-layer authentication decisions and other non-network security decisions that CWE-654 also covers are outside the clause's scope, mirroring the A.8.20/CWE-523 scope seam.
- CWE-693partialblocks — A.8.21 mandates authentication, encryption, network connection controls and monitoring for network services, which blocks exploitation of the network-service-related slice of CWE-693; the remainder is bounded and nameable — protection mechanism failures in application logic, sandboxing, signing, virtualization, and other non-network-service domains are outside the clause's scope.
- CWE-757partialblocks — A.8.21 specifies security features and technical parameters for network services (authentication, encryption, connection controls) and can require providers to mandate the strongest available algorithm with no fallback, which blocks downgrade exploitation where the provider complies; however, the weakness lives in the protocol's negotiation logic itself, which the clause's service-level agreement does not reach, leaving a nameable remainder in flawed implementations.
- CWE-918partialblocks — A.8.21's network connection controls, egress restrictions and monitoring of network service use can block some SSRF exploitation (e.g., preventing the server from reaching internal services, cloud metadata endpoints, or arbitrary ports), but the application-level URL-fetching vulnerability remains and can still be exploited against any destination the network permits.
- CWE-923partialblocks — A.8.21's requirements for authentication, network connection controls, and technical parameters for secured connections can block CWE-923 at the network-service level (e.g., requiring certificate validation or endpoint authentication as a service feature), but the majority of CWE-923 instances are application-level defects — SSL certificate validation bugs, hostname checks, cross-domain policies, Android intent origin checks — which sit in product code that A.8.21's network-service governance does not reach.
- CWE-924partialblocks — A.8.21 mandates security features for network services including encryption and connection controls; well implemented, these provide integrity protection over the network service portion of transit, masking the vulnerability during that portion — analogous to A.8.20 vs CWE-523 — but the product-level failure to enforce integrity remains, so end-to-end exploitation is not blocked.
- CWE-924partialprevents — A.8.21's guidance on security features of network services (encryption, connection controls, technical parameters for secured connections) requires integrity protection to be specified for network services, so a well-implemented clause prevents the weakness from arising in network service channels; the remainder is communication channels that are not network services (application-to-application channels, internal IPC, etc.) which the clause does not reach.
- CWE-941partialblocks — A.8.21's network connection controls, authentication requirements, and monitoring of network service use can block some exploitation paths for CWE-941 (e.g., egress controls preventing connections to unauthorized destinations, monitoring flagging anomalous destinations), but the underlying code defect that specifies the wrong destination remains, and the clause's primary focus is governing access to network services rather than addressing destination specification in outgoing communications.
Mitigated MITRE ATT&CK techniques (2989)
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
- T1021mostlyprevents — Requiring authentication, authorization, and connection controls for network services limits an adversary's ability to abuse remote services for lateral movement.
- T1040mostlyprevents — A.8.21 includes encryption, authentication, and network connection controls as required security features for network services, so where these are implemented (e.g., VPN, TLS) the bulk of sniffing is prevented; the same bounded remainder as A.8.24 applies — traffic volumes, headers, and any unencrypted service traffic still leak information.
- T1110.001mostlydetects — A.8.21 explicitly requires monitoring of the use of network services plus authentication/authorization procedures and anomalous access attributes, which surfaces password guessing attempts (especially on the listed management ports and protocols) once they occur; the remainder is non-monitored or out-of-scope services.
- T1133mostlyprevents — Defining allowed networks, authentication rules, and monitoring for external network services restricts adversaries from establishing persistent external remote access.
- T1669mostlyprevents — A.8.21 explicitly names wireless networks among the means of access it governs and mandates authentication, encryption, access controls, and monitoring for network services — which, when properly implemented, prevents open-Wi-Fi exploitation and unauthorized wireless access. The bounded remainder is the valid-accounts path (the clause requires authentication but cannot prevent use of legitimately held credentials) and implementation gaps where required measures are not fully enforced, mirroring the shape of A.8.5 vs T1110.001.
- T1001.001partialdetects — A.8.21 includes 'monitoring of the use of network services' (item g), which provides a vantage point on network traffic where junk-data C2 could be observed as anomalous protocol patterns, but the clause sets scope rather than mandating the deep protocol-analysis depth needed to reliably distinguish junk-padded C2 from legitimate traffic — and the technique is explicitly designed to defeat trivial decoding/analysis, so what is caught is a slice rather than the bulk.
- T1001.003partialdetects — A.8.21 includes 'monitoring of the use of network services' and rules on authentication/authorization for network access, which can surface unauthorized services or anomalous usage patterns, but the technique is specifically designed to blend C2 into legitimate traffic by faking TLS handshakes and manipulating HTTP headers/certificates — A.8.21 does not mandate the protocol-level fingerprinting or deep inspection needed to reliably catch impersonation, so detection is a slice rather than a bounded remainder of a working mechanism.
- T1011partialprevents — A.8.21 can establish rules on allowed networks/services (including wireless), require authentication, and monitor network service use, which constrains some alternative-medium exfiltration paths; however, T1011 is specifically chosen because the medium (Bluetooth, cellular, RF) is not routed through the enterprise network and is largely outside the scope of network service security controls, so a meaningful slice of the technique remains reachable.
- T1011.001partialprevents — A.8.21 requires rules on which networks and network services may be accessed, authentication and authorization for network service use, monitoring of network service use, and procedures to restrict access — and explicitly names wireless as a covered means of access. Bluetooth falls within that scope, so a properly implemented A.8.21 can prohibit or restrict Bluetooth use, require authentication for pairing, and monitor Bluetooth traffic. The extent is partial because the clause is governance-oriented: it requires rules to be formulated and implemented but does not itself mandate the specific technical controls (Bluetooth radio disabling, pairing controls, Bluetooth-specific DLP) that actually stop the exfiltration, and Bluetooth radios are commonly enabled by default on endpoints the clause does not directly reach.
- T1018partialprevents — A.8.21's network connection controls, segmentation, and access restrictions can limit the scope of what an adversary can discover from a compromised host, but the technique uses native OS utilities (ping, ARP, net view, host files) that do not require accessing a governed network service, so within an accessible segment the discovery still runs.
- T1020partialprevents — A.8.21 can restrict which networks and services may be accessed, require authentication/authorization for network service use, and mandate encryption of network connections — closing some exfiltration paths. But automated exfiltration typically rides on allowed channels (C2, alternative protocols) the control does not block, and the clause governs how services are used rather than what leaves through them, leaving a substantial remainder.
- T1020.001partialprevents — A.8.21 requires authorization procedures, network management controls, authentication, and monitoring of network service use — all of which constrain who can configure traffic mirroring on network devices or cloud platforms. However, it is a governance control that mandates measures rather than implementing them, and firmware-level compromises (ROMMONkit, Patch System Image) that modify the device itself bypass the governance layer entirely, leaving a named remainder.
- T1021.001partialprevents — A.8.21 requires network service rules covering authentication, authorization, and access restrictions, which can block RDP at the network layer, but the control governs service provision and network-level access rather than host-level RDP enablement, leaving the technique reachable where RDP is permitted or host configuration is weak.
- T1021.002partialprevents — A.8.21 requires authorization procedures for who may access which networks and services, network connection controls, and procedures to restrict access to network services — these can constrain which accounts reach admin shares and from where. But the technique uses valid accounts that satisfy A.8.21's authentication requirements, and admin shares are by design accessible to administrators, so the clause's framework reduces rather than eliminates the attack surface.
- T1021.003partialprevents — A.8.21's authorization procedures (item c) and network management/technological controls (item d) can restrict which accounts may invoke DCOM and which network connections permit RPC/DCOM traffic, but the clause governs network services generally rather than targeting DCOM specifically, so only a slice of the technique is prevented.
- T1021.004partialprevents — A.8.21 governs network service use and can require restrictions that constrain SSH abuse — key-only authentication, source-network allowlists, VPN prerequisites, time/location attributes, and monitoring that feeds blocking — but it is a governance clause that specifies requirements rather than implementing technical controls, and it cannot prevent an adversary with valid, authorized credentials from using SSH within the rules it sets.
- T1021.005partialprevents — A.8.21 requires authentication, authorization, and network connection controls for network services, which can block unauthorized VNC access, but VNC can be configured with its own credentials or left with weak defaults, leaving a reachable remainder.
- T1021.006partialprevents — A.8.21 requires network service rules covering authentication, authorization, and technological controls to protect access, which can restrict or block WinRM usage, but the control sets requirements rather than mandating a specific blocking mechanism, leaving room for implementations that permit WinRM under certain conditions.
- T1021.008partialprevents — A.8.21 governs network services and can require authorization procedures, access restrictions, and technical parameters for services like cloud VM direct connections, which constrains who may use them; however, the technique specifically leverages valid accounts that pass authentication and authorization, so the control's framework permits rather than blocks the access when credentials are valid.
- T1029partialprevents — A.8.21 restricts allowed network services, access rules, and time/location attributes, which can block some scheduled exfiltration paths, but transfers scheduled over approved services during normal hours still succeed.
- T1039partialprevents — A.8.21's authorization procedures and authentication requirements for network services can restrict which users access network shares, but the control is procedural/governance aimed at network service providers and broad network access rules rather than file-share permissions specifically (A.8.3's domain), and the adversary is already on a system with some level of access.
- T1040partialdetects — A.8.21's monitoring of network service use (clause g) and network management controls (clause d) can surface some indicators of sniffing — anomalous access patterns, unauthorized endpoints, unusual span-port or promiscuous-mode activity — but the clause's primary thrust is preventive (encryption, authentication, access controls) and its monitoring scope is service-use rather than passive capture, so detection is a slice rather than the bulk.
- T1046partialdetects — A.8.21's clause (g) on monitoring the use of network services can surface anomalous scanning activity against the organization's services, but the control's primary purpose is access rules and provider security rather than reconnaissance detection, and the monitoring scope is 'use of services' rather than external scan detection.
- T1046partialprevents — A.8.21's network connection controls, authentication requirements, access restrictions, and monitoring can reduce the discoverable surface and block some scanning attempts, but the technique still runs against services that must remain reachable — discovery of an exposed service is not the same as access to it, and the clause secures services rather than removing them.
- T1048partialdetects — A.8.21 requires that rules be formulated and implemented covering monitoring of network service use (g), technical parameters for secured connections (b), and procedures to restrict access to network services (d); when implemented, these surface anomalous use of alternative protocols for exfiltration, but the control itself is a governance/management clause that mandates the capability rather than performing the detection, so what is caught is a slice of what monitoring of network service use would see, not the bulk.
- T1048partialprevents — A.8.21 can restrict which network services and protocols are accessible (e.g., block outbound FTP/SMB), require authentication, and implement network connection controls, which prevents some alternative-protocol exfiltration; however, many of the protocols T1048 uses (HTTPS, DNS, SMTP) are typically permitted for business reasons, and the clause governs network service access rather than inspecting exfiltrated content, so a bounded remainder of exfiltration over allowed channels survives.
- T1048.001partialprevents — A.8.21 can restrict which network services and destinations are reachable, require authentication/authorization, and mandate monitoring of network service use, which constrains the paths available for exfiltration; however, the technique uses common allowed protocols (HTTP, FTP, HTTPS) with symmetric encryption that defeats content inspection, so when the adversary operates within authorized access the control cannot prevent the data from leaving.
- T1048.002partialprevents — A.8.21 establishes rules on allowed networks/services, authentication, authorization, monitoring of network service use, and technical parameters for secured connections, which can restrict destinations and detect anomalous use; however, T1048.002 uses legitimate asymmetrically encrypted protocols (HTTPS/TLS) baked into the protocol itself, so governance-level controls over network service use cannot fully prevent the exfiltration channel.
- T1048.003partialdetects — A.8.21 includes 'monitoring of the use of network services' and mandates security features such as encryption and network connection controls, so unencrypted exfiltration traffic can be surfaced where the implementation monitors for compliance with those requirements; however, the clause's monitoring is primarily about access patterns (who, when, where, which service) rather than content inspection, and the scope of what is actually monitored is set by implementation choice, leaving a bounded remainder.
- T1048.003partialprevents — A.8.21 mandates encryption, network connection controls, and monitoring of network services, which constrains exfiltration over unencrypted protocols where those services are in scope; however, the clause governs the relationship with network service providers and sets requirements rather than implementing direct technical controls, leaving gaps for local network traffic, protocols outside service agreements, and implementation shortfalls.
- T1049partialdetects — A.8.21's monitoring of network service usage (item g) and network management controls (item d) can surface network-based variants of T1049 where the adversary queries remote systems for connection information, but the core of the technique—local enumeration via netstat, lsof, who, w, and similar commands on a compromised host—falls outside the clause's monitoring scope, which is oriented toward service access patterns rather than local reconnaissance commands.
- T1059.008partialprevents — A.8.21's authentication, authorization, and network connection control requirements for network services can prevent unauthorized access to network device CLIs, but cannot prevent authorized users from abusing their legitimate access, does not address physical console access, and is a governance-level clause that mandates requirements rather than implementing the technical controls itself.
- T1059.009partialprevents — A.8.21's authentication requirements (b) and authorization procedures (c) can prevent unauthorized access to cloud APIs as network services, but the technique's core is abuse by adversaries holding valid credentials (Application Access Tokens, Web Session Cookies) invoking legitimate API functions — the clause governs network-level access, not IAM/API-level permissions, so it does not reach the bulk of the technique.
- T1059.013partialprevents — A.8.21's requirements for authentication, authorization, network connection controls, and monitoring of network services constrain access to the Docker daemon API and Kubernetes API server over the network, but the technique also executes via local CLI tools (docker, kubectl) and locally-mounted Docker sockets that bypass network service controls entirely, and the clause sets requirements rather than mandating specific mechanisms.
- T1071partialprevents — A.8.21's authentication, authorization, and network connection controls can prevent unauthorized use of network services for C2 and restrict which services are reachable, but the technique uses application layer protocols (HTTP, DNS, SMTP, etc.) that are required for normal business operations, so the control governs access without eliminating the protocol surface the technique relies on.
- T1071.001partialdetects — A.8.21 explicitly mandates 'monitoring of the use of network services' (item g) and security features including authentication and connection controls, which provide some detection surface for anomalous web protocol use; however, the clause is primarily governance/procedural (specifying requirements, audit rights, and provider attestations) with monitoring as one of seven rules, and T1071.001's blending with normal traffic means what is caught is a slice of the technique rather than a bounded remainder of a working mechanism.
- T1071.001partialprevents — A.8.21 can require authentication, authorization, VPN use, egress controls, and proxying of network services, which constrain how web protocols are used and can block direct C2 channels, but HTTP/S is fundamental to most operations and the technique blends with legitimate traffic, so the control reduces rather than eliminates the attack surface.
- T1071.002partialprevents — A.8.21 can restrict which file transfer protocols are permitted (e.g., block unencrypted FTP, require FTPS), mandate authentication and authorization for their use, and set technical parameters for secured connections — all of which constrain T1071.002. But the technique's own description notes SMB, FTP, FTPS and TFTP are 'very common in environments' and often business-required, so the control governs and hardens their use rather than eliminating the protocols, leaving a substantial remainder where the technique still runs over legitimately permitted services.
- T1071.003partialprevents — A.8.21 provides the framework to restrict access to network services, implement network connection controls, and monitor use of network services — mechanisms that can prevent direct outbound SMTP/POP3/IMAP to non-corporate servers and constrain where mail protocol traffic is permitted. However, mail protocols are essential for business and cannot be blocked wholesale, and the clause sets requirements rather than mandating specific technical mechanisms, so abuse of legitimate mail services (including the corporate mail server itself) and traffic that blends with normal mail flows remains reachable.
- T1071.004partialdetects — A.8.21's requirement to monitor the use of network services (item g) and apply network management/technological controls to protect access to network connections (item d) can surface anomalous DNS usage patterns, but the clause is primarily a governance/requirements control governing the relationship with network service providers rather than a dedicated detection mechanism; DNS tunneling is specifically designed to blend with normal traffic, so what monitoring under this clause catches is a slice rather than the bulk.
- T1071.005partialprevents — A.8.21 can restrict which network services are accessible, require authentication for network service access, and implement network connection controls that block unauthorized pub/sub brokers — but it is a governance/management control that establishes rules rather than mandating specific technical blocks, and if the organization legitimately uses MQTT/XMPP/AMQP/STOMP for business purposes the technique still has a channel, with adversaries specifically designed to blend into that allowed traffic.
- T1072partialprevents — A.8.21's authentication, authorization and monitoring requirements for network services can prevent unauthorized access to software deployment tools, but the technique commonly runs with valid administrative credentials or abuses legitimate functionality, which A.8.21's general network-service requirements do not address.
- T1078partialprevents — A.8.21 constrains access to network services through authentication requirements, authorization procedures, time/location attributes, and means-of-access controls (e.g., VPN), which can block some abuses of valid credentials (e.g., where MFA or conditional access is not satisfied, or where the account is not authorized for the service), but it cannot prevent the technique when the adversary holds credentials that satisfy the configured authentication and authorization — the accounts are valid and pass the controls the clause mandates.
- T1078.002partialprevents — A.8.21's authentication requirements, authorization procedures, and monitoring of network service use can constrain abuse of domain accounts against network services (e.g., requiring MFA or specific auth mechanisms for service access), but it does not prevent the technique broadly: it only covers network services, not all domain account usage, and does not stop credential theft or abuse of valid credentials for non-network-service purposes.
- T1080partialprevents — A.8.21's authorization procedures and access restrictions to network services can limit who can write to shared network drives, which constrains the tainting surface, but the control addresses the network service layer (authentication, connection security, service-level agreements) rather than content integrity on shared storage, so it does not reach the bulk of the technique.
- T1090partialdetects — A.8.21's item (g) requires monitoring of the use of network services, and its security features include network connection controls and procedures to restrict access, which can surface proxy usage patterns (unexpected intermediaries, anomalous routing, CDN abuse) on the organization's estate; however, the clause's primary purpose is governing the provider relationship and setting usage rules, so detection of T1090 is a side effect of monitoring scope rather than a dedicated mechanism, and proxies that operate entirely outside monitored network services remain unseen.
- T1090partialprevents — A.8.21's technical parameters, authentication requirements, network connection controls, and monitoring of network service use can constrain some proxy usage (e.g., requiring authenticated egress, restricting which networks/services may be accessed, monitoring for anomalous flows), but the technique fundamentally operates on adversary-controlled infrastructure or compromised hosts that the clause does not reach, so what is prevented is a slice rather than the technique.
- T1090.002partialprevents — A.8.21's rules on allowed networks/services, authentication for network service access, and procedures to restrict access can block arbitrary outbound connections to external proxies via egress controls, but the technique is explicitly designed to ride over existing trusted communications paths and use low-likelihood-investigation infrastructure, so the clause's requirements (set rather than mandate universal blocking) leave a meaningful slice where the proxy sits on an allowed path and the technique still runs.
- T1090.003partialdetects — A.8.21 includes 'monitoring of the use of network services' and network connection controls, and multi-hop proxy traffic does traverse the organization's network at the last hop, so there is something to observe; however, the technique is specifically designed to defeat source attribution, and A.8.21's monitoring is general service-use monitoring rather than proxy-chain analysis, so only a slice of the activity is surfaced.
- T1090.004partialprevents — A.8.21's clauses on monitoring network service use (g), technical parameters for secured connections (b), and procedures to restrict access to network services (d) provide mechanisms that can be applied against domain fronting — e.g., restricting which CDNs may be used, inspecting outbound traffic for SNI/Host anomalies, or enforcing network connection controls — but the control is general network-service security and does not specifically mandate the deep inspection or CDN-egress restrictions needed to close the technique, so what is prevented is a slice rather than the bulk.
- T1095partialdetects — A.8.21 includes 'monitoring of the use of network services' and security features like network connection controls, which can surface non-application layer protocol abuse, but the clause is primarily a governance/requirements control over providers and rules rather than a detection mechanism; non-application layer protocols are explicitly noted as less commonly monitored, and the clause does not mandate the depth of instrumentation needed to catch them reliably.
- T1095partialprevents — A.8.21 can require network connection controls, protocol restrictions, and monitoring that constrain which non-application layer protocols are permitted on the network, but it is a governance clause that mandates these measures rather than implementing them itself; ICMP is required for IP-compatible hosts and VMCI traffic is local to the hypervisor, so the protocols T1095 most exploits are difficult to block without breaking functionality, leaving a bounded remainder.
- T1102partialprevents — A.8.21 provides mechanisms (allowlisting of permitted networks/services, technical controls to protect network connections, procedures to restrict access to network services) that can block use of some web services for C2, but T1102 specifically leverages legitimate, commonly-used services (Google, Microsoft, Twitter) that organizations typically must allow for business, so the control cannot prevent the technique without disrupting operations.
- T1102.002partialprevents — A.8.21 can restrict which networks and services are allowed (item a), implement network connection controls, and enforce authentication/authorization on network service use, which constrains some bidirectional C2 over legitimate web services; however, the technique specifically leverages popular services (Google, Twitter, etc.) that organizations typically must allow for business reasons, the traffic is SSL/TLS encrypted and indistinguishable from normal use, and many alternative services exist, so the control cannot fully prevent the technique.
- T1102.003partialprevents — A.8.21's clauses on restricting which networks/services may be accessed, network connection controls, and procedures to restrict access to network services can be implemented to block or constrain use of specific external web services, but the technique is specifically designed to ride legitimate, commonly-allowed services (Google, Twitter, etc.) over standard SSL/TLS, so wholesale blocking is impractical and the control cannot reach full prevention.
- T1105partialprevents — A.8.21's restrictions on which networks and services may be accessed, its authentication and authorization requirements for network services, and its network management/connection controls can block some ingress tool transfer vectors (e.g., FTP/TFTP, unauthenticated services, disallowed external endpoints), but the technique's primary vectors — HTTPS to external servers, native utilities like curl/wget/PowerShell, and cloud sync services — typically traverse allowed network services, so the clause constrains rather than closes the technique.
- T1110partialprevents — A.8.21 requires authentication, authorization, and network connection controls for network services, which can block brute force attempts at the network perimeter, but the technique also operates offline against stolen hashes and through services not covered by these rules, leaving a substantial remainder.
- T1110.001partialprevents — A.8.21 mandates authentication requirements, access restrictions, network connection controls, and procedures to restrict access to network services — the framework that, when implemented, prevents password guessing against those services — but it is a governance/requirements control that specifies what providers and the organization must put in place rather than implementing the technical mechanisms itself, so the actual prevention depends on downstream implementations (which A.8.5 covers more directly at `mostly`).
- T1110.003partialprevents — A.8.21 requires authentication mechanisms (including MFA-capable options), access restrictions on which network services are exposed, and monitoring of network service use — all of which can prevent password spraying from succeeding. However, it is a governance/requirements control that mandates these measures rather than implementing them, and password spraying can still succeed against services that don't fully implement the required controls or that use legacy authentication exempted from the requirements.
- T1110.004partialdetects — A.8.21 mandates monitoring of network service use and specifies authentication requirements and technical parameters that frame what anomalous authentication patterns look like, contributing to detection of credential stuffing; however, the control is a requirements/governance clause that specifies monitoring should be in place rather than performing the monitoring itself, so detection is partial and the actual surfacing is performed by A.8.16 and similar monitoring controls.
- T1110.004partialprevents — A.8.21 requires authentication, authorization, and network connection controls for network services, which can block credential stuffing attempts at the perimeter, but the technique targets application-layer services (SSO, cloud apps, email) where the control's network-level measures do not reach, leaving a substantial remainder.
- T1114partialprevents — A.8.21 governs network service access — authentication, authorization, encryption, and access restrictions for services like email — which prevents the network-access portion of T1114 (unauthenticated or unauthorized access to mail servers, eavesdropping on email traffic). It does not prevent collection from local email clients (PSTs, OSTs) once a host is compromised, nor collection via authorized but compromised accounts or stolen API tokens, so the prevention is partial.
- T1114.002partialprevents — A.8.21's authentication requirements, authorization procedures, and access restrictions can prevent unauthorized access to email services, but T1114.002 explicitly leverages valid credentials or access tokens to interact with Exchange/O365/Gmail, so once credentials are valid the technique runs against services the account is authorized to use — the control prevents a slice (unauthorized access) but not the bulk of the technique.
- T1135partialprevents — A.8.21's authentication requirements, authorization procedures, and network connection controls can block unauthenticated and unauthorized share enumeration (e.g., null session discovery), but the technique is fundamentally a discovery action that authenticated, authorized users can still perform against shares they have access to, so the control prevents a slice rather than the bulk.
- T1187partialdetects — A.8.21 requires monitoring of network service use (item g) and technical controls protecting network connections (item d), and forced authentication generates observable outbound SMB traffic to unauthorized destinations that such monitoring could surface — but the control is primarily governance-oriented, the monitoring scope is general rather than targeted at this technique, and the actual detection depends on implementation specifics.
- T1187partialprevents — A.8.21's rules on which networks and services may be accessed, authentication requirements, and network management/technological controls to protect network connections can partially constrain forced authentication by restricting outbound SMB/WebDAV to untrusted destinations and mandating authentication mechanisms (e.g., SMB signing), but the control is fundamentally about governing network service use and provider relationships rather than directly addressing the client-side auto-authentication behavior that T1187 exploits.
- T1190partialprevents — A.8.21 prevents some exploitation by restricting which network services are exposed, requiring authentication and authorization, and applying network connection controls — but T1190 by definition targets applications that must remain publicly accessible, and the clause addresses access-layer security rather than application-layer vulnerabilities (bugs, code flaws, patching), so a bounded slice of the technique is prevented and the remainder is not.
- T1199partialprevents — A.8.21's requirements for authentication, authorization scoping, and monitoring of network service use can prevent successful abuse of a compromised third party's access in many cases, but it does not address the third party's own security posture, the vetting of providers before access is granted, or the initial compromise of the third party — so the technique can still be attempted and succeed where those gaps remain.
- T1200partialprevents — A.8.21 mandates network connection controls, authentication for network services, and monitoring of network service use, which can prevent the network-based variants of T1200 (rogue access points, network taps) by blocking unauthorized devices from joining the network; however, the technique also encompasses non-network hardware additions such as USB keystroke injection and DMA attacks, which fall outside the clause's scope of network service security.
- T1210partialdetects — A.8.21 includes 'monitoring of the use of network services' (item g) among its required rules, which can surface anomalous connections and unexpected service usage indicative of remote service exploitation, but the clause is primarily a governance/management control about establishing security requirements and ensuring providers implement them — detection is a secondary aspect, not the primary mechanism.
- T1210partialprevents — A.8.21's network connection controls, access restrictions, and segmentation can prevent adversaries from reaching vulnerable remote services, but it does not address the underlying software vulnerabilities themselves (A.8.8's domain), so services that must remain accessible for business reasons can still be exploited.
- T1213partialprevents — A.8.21's authentication, authorization, and network access controls apply to repositories as network services and prevent unauthorized network-level access, but T1213 commonly succeeds via legitimate credentials, application-layer misconfigurations (overly-broad sharing, public access), and authorized users mining data they can already see — none of which A.8.21 addresses.
- T1213.003partialprevents — A.8.21's authentication, authorization, and monitoring requirements for network services can prevent unauthorized access to code repositories accessed as network services (web/git), but the technique also covers scenarios where adversaries already have valid access or where the repository's content-level access controls (not network service security) are the actual barrier.
- T1213.006partialprevents — A.8.21's authentication, authorization, and network access controls can prevent unauthorized network access to databases, but the technique still succeeds when adversaries use valid credentials, access through authorized channels, or reach cloud/SaaS databases via legitimate APIs that fall outside traditional network service controls.
- T1219partialdetects — A.8.21's item (g) requires monitoring of the use of network services, and remote access tools are network services, so monitoring their use can surface T1219 activity; however, monitoring is one of seven sub-items in a control primarily oriented toward security requirements and access controls, and the clause does not mandate the behavioral depth needed to reliably distinguish legitimate remote access tool use from adversary abuse.
- T1219partialprevents — A.8.21's authentication, authorization, network connection controls, and monitoring of network services can prevent unauthorized remote access channels, but T1219's bulk — legitimate tools already permitted, built-in features (Chrome Remote Desktop, EDR response features), and reverse/outbound connections from compromised hosts — sits outside what governing network service access can stop.
- T1219.001partialdetects — A.8.21's clause (g) requires monitoring of network service use and (d) requires network management and technological controls to protect access, which can surface unauthorized tunneling protocols, access from unusual locations/times, or use outside approved means — but the technique is specifically designed to blend in with legitimate developer workflows using legitimate authentication (e.g., GitHub) and standard protocols, and A.8.21 is primarily a governance/management control rather than a technical detection mechanism, so what it catches is a slice of policy-violating use rather than the technique itself.
- T1219.001partialprevents — A.8.21's framework for restricting network service access, requiring authorization, and implementing network connection controls could be configured to block IDE tunneling applications or restrict their use, but the technique is specifically designed to use legitimate authentication (e.g., GitHub accounts) and blend in with normal developer workflows, so the control's standard mechanisms do not inherently prevent it.
- T1219.002partialdetects — A.8.21 mandates monitoring of network service use (item g) and includes authentication/authorization requirements that generate logs, which could surface remote desktop software abuse; however, it is a governance/requirements control that sets the framework rather than implementing detection itself, so the actual detection depends on the monitoring tools deployed and their scope.
- T1219.002partialprevents — A.8.21 can restrict which network services are allowed (a), require authentication (b) and authorization (c), apply network controls (d), and monitor use (g), which constrains how remote desktop software is accessed and used; however, the technique explicitly leverages legitimate RMM tools commonly permitted in environments, and A.8.21 governs rather than blocks, so where such software is allowed for legitimate support purposes the technique still operates within those authorized channels.
- T1219.003partialprevents — A.8.21 governs network services including authentication, authorization, allowed services, and monitoring of network service use, which constrains unauthorized KVM devices on the network; however, the technique involves physical hardware installation and post-compromise deployment that bypasses network service provisioning, so the control cannot reach the bulk of the technique's surface.
- T1491.002partialprevents — A.8.21's authentication, encryption, network connection controls, authorization procedures, and access restrictions reduce vectors for unauthorized access to externally-facing systems, but defacement typically exploits application-layer vulnerabilities, compromised credentials, or hosting/DNS infrastructure outside the clause's scope, so prevention is partial.
- T1496partialdetects — A.8.21's monitoring of network service usage (item g) and technical parameters for secured connections can surface bandwidth-based hijacking such as proxyjacking and unusual outbound traffic, but the control's scope is network services — compute-resource hijacking (cryptojacking, container cryptojacking) and abuse of cloud messaging quotas fall outside its monitoring vantage point.
- T1496partialprevents — A.8.21's authentication, authorization, and network access controls can prevent some vectors of resource hijacking — particularly unauthorized access to cloud services for compute abuse and network bandwidth selling — but it does not address endpoint malware-based hijacking such as cryptomining, which is a primary vector for T1496 and falls outside the clause's scope of network service security.
- T1496.002partialdetects — A.8.21(g) requires monitoring of the use of network services, which could surface anomalous bandwidth consumption patterns indicative of bandwidth hijacking, but the clause is primarily a governance/management control governing service-provider relationships and access rules rather than a technical monitoring control, so detection is incidental rather than designed and would not reliably catch proxyjacking, torrent seeding, or scanning traffic that blends with normal use.
- T1496.004partialdetects — A.8.21's clause (g) on monitoring the use of network services and its security-feature requirements (authentication, connection controls) can surface some indicators of T1496.004 — unusual API volumes, anomalous access patterns, unexpected service enablement — but the technique is specifically designed to blend in with legitimate SaaS use (often services the victim already uses), and A.8.21's monitoring is framed around service-provider compliance and rule adherence rather than deep behavioral anomaly detection, so what is caught is a slice rather than the bulk.
- T1496.004partialprevents — A.8.21 can restrict which network services are accessible, require authentication/authorization for their use, and apply technical parameters for secured connections — preventing unauthorized SaaS access and limiting which services can be enabled. But the technique explicitly leverages services the victim is already using with valid (compromised) credentials, and A.8.21's rules on access and authorization do not stop abuse of legitimately authorized services by an adversary who holds valid tokens.
- T1499.003partialprevents — A.8.21's provisions for access restrictions, monitoring of network service use, and procedures to limit access to network services or applications can constrain some application exhaustion floods (e.g., rate limiting, IP-based access controls), but the clause governs network service security governance rather than mandating application-layer resource protection, so resource-intensive features can still be exhausted by requests that pass the network-service gates.
- T1499.004partialprevents — A.8.21's network access controls, authentication requirements, and restrictions on which network services may be accessed can prevent some exploitation paths to vulnerable applications, but the control addresses network service security rather than software vulnerability exploitation itself, so it blocks a slice of attack paths rather than the technique's core mechanism.
- T1530partialprevents — A.8.21 mandates authentication, authorization, network connection controls, and access restrictions for network services including cloud storage, which addresses the access-control surface of T1530; however, the technique's primary vectors — misconfigured storage permissions and stolen/leaked credentials — remain reachable because the clause sets requirements and provider agreements rather than guaranteeing correct configuration, and the technique description explicitly notes misconfigurations as a common problem.
- T1537partialdetects — A.8.21's monitoring of network service usage (clause g) and its authentication/authorization/access-attribute requirements (clauses b, c, f) can surface some unauthorized cloud data movements, but T1537 is explicitly designed to evade traditional network-service monitoring by using cloud provider APIs and the provider's internal address space, so what is caught is a slice rather than the bulk.
- T1538partialprevents — A.8.21 mandates authentication requirements (potentially MFA), authorization procedures, and access controls based on time/location/attributes for network services including cloud dashboards; if applied to cloud services these would block credential-only access. But the control is a management/governance clause that requires measures to be identified and implemented rather than implementing them itself, and stolen credentials may include session tokens or be used from trusted locations, so the prevention is a mandated capability rather than a direct mechanism.
- T1539partialdetects — A.8.21 includes 'monitoring of the use of network services' as one of its rule areas, which can surface anomalous network patterns consistent with AitM cookie theft (e.g., Evilginx2-style proxies), but the control's primary purpose is preventive — establishing authentication, authorization and service-level rules — and the monitoring bullet is one item among many; disk-based, in-memory, and JavaScript-injection cookie theft vectors fall outside what network-service-use monitoring observes.
- T1539partialprevents — A.8.21's encryption and network connection controls prevent cookie theft from network traffic (e.g., unencrypted sniffing) and can constrain AitM proxies via network controls, but it does not address cookie theft from local disk, browser process memory, or malicious JavaScript injection — vectors the technique explicitly enumerates and which require endpoint or web-application controls.
- T1542.004partialprevents — A.8.21's authentication, authorization, and network connection controls can restrict remote TFTP-based ROMMON updates and limit who may push firmware to network devices, but the clause does not address firmware integrity verification, boot-loader signing, or local physical access — the vectors that make ROMMONkit difficult to detect and that fall outside 'security of network services' as the clause frames it.
- T1542.005partialprevents — A.8.21 requires authentication, authorization, and network connection controls that can restrict which TFTP servers a device is permitted to boot from, but the technique abuses a standard administrative function and a device already compromised to change its boot configuration may bypass those controls, so the prevention is a slice rather than a closure.
- T1552.005partialprevents — A.8.21's requirements for network connection controls, access restrictions, and monitoring of network services can partially constrain metadata API access (e.g., firewall rules, authorization procedures for which services may reach 169.254.169.254), but the technique exploits a designed feature of cloud infrastructure accessible from any process on the instance, and the SSRF vector is an application-layer vulnerability outside A.8.21's scope, so prevention is bounded.
- T1552.007partialprevents — A.8.21 mandates authentication, authorization, and monitoring for network services, which includes container APIs; properly implemented, these requirements prevent unauthorized access to Docker/Kubernetes APIs, but the technique can still succeed via authorized access (e.g., a pod's service account with legitimate permissions retrieving secrets), so the control addresses a slice rather than the bulk.
- T1557partialdetects — A.8.21 includes 'monitoring of the use of network services' (item g) and requires technical parameters and network connection controls that can surface anomalous traffic patterns indicative of AiTM positioning, but the control's primary purpose is establishing security requirements for network services rather than mandating detection instrumentation, and the depth of monitoring is implementation-dependent — AiTM via ARP/DNS/LLMNR poisoning can be stealthy and may evade basic usage monitoring.
- T1557partialprevents — A.8.21 mandates authentication, encryption, and network connection controls that can prevent some AiTM vectors (e.g., enforcing VPNs, authenticated connections), but the technique exploits protocol-level features like ARP and DNS that operate below the application-layer security the control primarily governs, leaving a substantial reachable surface.
- T1557.002partialdetects — A.8.21 includes 'monitoring of the use of network services' (clause g) and technical controls to protect network connections, which can surface anomalous ARP behavior, but the clause sets requirements and scope rather than mandating the specific instrumentation (DAI, ARP-aware IDS) needed to reliably detect cache poisoning; what is caught is a slice determined by implementation choices, not a bounded remainder of a mechanism that mostly works.
- T1557.004partialprevents — A.8.21 can require VPN usage, certificate-based authentication, and network connection controls that would prevent the evil twin technique from succeeding (e.g., a device validating 802.1X certs will not associate with a rogue AP, and VPN-wrapped traffic is unreadable through the rogue), but the technique itself — the adversary hosting a rogue AP — is not prevented, and the clause sets requirements rather than mandating universal mechanisms, so users on networks without these controls remain reachable.
- T1558.001partialdetects — A.8.21's clause (g) on monitoring the use of network services and its consideration of authentication security features could encompass monitoring Kerberos authentication traffic, but the control is a general network-services provision clause that does not specifically mandate the Kerberos-level telemetry (ticket encryption types, lifetimes, TGS request anomalies) needed to reliably surface a forged golden ticket, so detection is a slice of what the monitoring scope could include rather than a bounded remainder of a mechanism that mostly works.
- T1565.002partialprevents — A.8.21 requires encryption and connection controls for network services, which prevents in-transit manipulation over the network where integrity-protected encryption (TLS/IPsec) is in force; however, the technique also operates between system processes on a single host and at endpoints before encryption/after decryption, both outside A.8.21's network-services scope, and the clause calls for encryption without explicitly mandating integrity protection.
- T1566.003partialprevents — A.8.21 can require and enforce restrictions on which network services employees may access from corporate devices (e.g., blocking personal webmail and social media), which prevents the technique from succeeding on managed endpoints; however, the technique is explicitly designed to use third-party services outside enterprise control and can still land on personal devices or via networks A.8.21 does not govern, so the prevention is a slice rather than a closure.
- T1567partialdetects — A.8.21 includes 'monitoring of the use of network services' as one of its mandated rules and requires technical controls to protect network connections, so some detection of anomalous outbound traffic to web services is in scope; however, the control is primarily a governance clause establishing rules, authentication, authorization and security features for network services, and the monitoring element is one of seven listed items whose depth is set by implementation rather than mandated, so detection of T1567's legitimate-looking HTTPS exfiltration is a slice rather than a bounded remainder of a working mechanism.
- T1567partialprevents — A.8.21 provides rules on which networks and services may be accessed, authorization procedures, and technical controls to restrict access to network services — mechanisms that can block specific web-based exfiltration channels. However, T1567 specifically leverages legitimate, commonly-used web services that are typically already permitted for business operations, and A.8.21 sets a governance framework rather than mandating blocks on particular services, so the prevention is a slice of what the technique's design exploits.
- T1567.002partialdetects — A.8.21 mandates 'monitoring of the use of network services' (item g) as one of its required rules, and exfiltration to cloud storage traverses the network as use of a network service, so the mandated monitoring could surface it; however, A.8.21 is not itself the monitoring system — it requires the capability to be implemented, and the scope/depth of that monitoring is not specified, so detection is a slice of the control's requirements rather than a direct mechanism.
- T1567.002partialprevents — A.8.21 mandates rules on which network services may be accessed, authorization procedures, technical controls to protect access to network connections, and procedures to restrict access to network services — all of which can be applied to block or constrain cloud storage services and thus prevent some T1567.002 exfiltration. However, the clause is governance-level (it requires implementation rather than implementing the controls itself), cloud storage is often legitimately needed for business so it cannot be fully blocked, and adversaries can use many services and channels, leaving a substantial remainder that the mandated controls cannot reach.
- T1567.003partialprevents — A.8.21 requires rules on which networks and network services may be accessed and procedures to restrict access to network services or applications where necessary, which can block text storage sites; however, the control is governance-level — it mandates the rules and their implementation but does not itself provide the technical enforcement (proxy filtering, content inspection, DLP), and authentication/authorization requirements do not stop an already-authenticated adversary from reaching an allowed service.
- T1567.004partialdetects — A.8.21 includes 'monitoring of the use of network services' (item g) and network connection controls, giving a detection surface over the HTTPS traffic webhook exfiltration rides on; however, the technique is specifically designed to blend in with normal traffic to legitimate SaaS endpoints, and A.8.21 mandates general monitoring of service use rather than the depth of content/behavioral analysis needed to reliably distinguish staged exfiltration from ordinary webhook traffic.
- T1568partialprevents — A.8.21's network connection controls, egress restrictions, and monitoring of network service use can constrain some dynamic resolution attempts (e.g., blocking unauthorized DNS, restricting outbound to non-approved services), but the control addresses the organization's use of network services rather than malware-driven C2 resolution, and T1568 is specifically designed to evade these network-layer controls.
- T1568.002partialdetects — A.8.21's clause (g) on 'monitoring of the use of network services' provides a surface where DGA-generated DNS queries would be visible, but the clause's focus is on access governance and service-provider security rather than mandating the statistical/ML analysis of domain-name patterns needed to reliably distinguish DGA traffic from legitimate use; DGAs are specifically designed to evade standard network monitoring.
- T1570partialdetects — A.8.21's point (g) requires monitoring of network service use, and lateral tool transfer operates over network services (SMB, RDP, SSH, etc.), so monitoring can surface anomalous file transfers and inter-system connections; however, the clause's primary focus is governing access to network services and security features rather than deep attack detection, and the monitoring scope is set by implementation rather than mandated to cover lateral movement specifically.
- T1570partialprevents — A.8.21 establishes authentication, authorization, network connection controls, and access restrictions for network services, which can constrain which services are reachable and require credentials for access — but T1570 operates over legitimate, authenticated channels (SMB, RDP, scp, etc.) between already-compromised systems, so the control's framework of rules and provider-level requirements cannot directly stop tool transfer once an adversary holds valid credentials and uses standard protocols.
- T1571partialprevents — A.8.21 mandates network connection controls, technical parameters for secured connections, monitoring of network service use, and procedures restricting access — all of which, when implemented, constrain non-standard port pairings. But the control is governance-level: it requires that rules and technical controls be in place, it does not itself implement the filtering or port restrictions that actually stop T1571. The prevention is real but indirect, and depends on the mandated controls being competently deployed.
- T1572partialdetects — A.8.21's monitoring of network service use (item g) and network management controls (item d) can surface tunneling that uses unauthorized services or produces anomalous access patterns, but the clause monitors access (who/what/when/where) rather than traffic content, so encapsulation within allowed protocols — the core of T1572 — is largely outside its detection scope.
- T1572partialprevents — A.8.21 mandates network connection controls, authentication for network services, authorization procedures, and access-restriction procedures that can block unauthorized tunneling, but it is a governance/management control that establishes requirements rather than directly implementing the technical controls; tunneling over allowed protocols (HTTPS, DNS, SSH) remains possible and the actual prevention depends on the depth of the technical controls the clause mandates.
- T1595.003partialprevents — A.8.21's clauses on restricting access to network services and applications, authentication requirements, and network connection controls can prevent wordlist scans from succeeding by removing discoverable content, gating paths behind authentication, and constraining what is exposed — but the technique still runs (probes still arrive at the perimeter), and publicly required services (web servers, cloud buckets) remain enumerable, so prevention is a slice rather than a closure.
- T1599partialprevents — A.8.21 mandates security features like authentication, encryption, and network connection controls that harden the boundary device against compromise, but it does not guarantee the device cannot be taken over, so the technique is constrained rather than eliminated.
- T1599.001partialdetects — A.8.21 includes monitoring of network service use (item g) and network management/technological controls (item d), which could surface anomalous traffic patterns or routing resulting from NAT modifications, but the clause's scope is governing network service security rather than detecting tampering with network device configurations themselves, so detection is incidental rather than primary.
- T1602partialprevents — A.8.21 mandates security features for network services — including authentication, encryption, access controls, and monitoring — and configuration repositories are network services, so the right mechanisms are named; but the clause is a governance/oversight control that requires the organization to identify and ensure implementation rather than mandating specific technical controls, and some configuration data is intentionally accessible to administrators, so the prevention is a slice rather than the bulk.
- T1602.002partialprevents — A.8.21 requires authentication, authorization, and monitoring for network services, which can prevent unauthorized access to network device configurations via SNMP/SMI when properly implemented, but it does not address protocol-level vulnerabilities (e.g., Smart Install's lack of authentication), credential compromise, or the specific configuration of network device management interfaces, and the technique targets network devices where these controls are often weakly enforced.
- T1609partialprevents — A.8.21 requires authentication, authorization, and network access controls for network services, and container administration services (Docker daemon, Kubernetes API, kubelet) are network services whose APIs would be covered — proper implementation prevents unauthorized remote execution. But the control is a requirements clause that depends on correct implementation, legitimate administrative access must be permitted (and can be abused with compromised credentials), and it does not address the container runtime or the commands themselves once access is gained.
- T1612partialprevents — A.8.21's authentication, authorization, and monitoring requirements for network services can block unauthorized access to the Docker API (a network service), but it is a general framework that does not specifically address container build operations, and the technique can still execute if the adversary holds legitimate access or compromised credentials to a user permitted to use the build API.
- T1613partialprevents — A.8.21's authentication, authorization and monitoring requirements for network services apply to container APIs (Docker/Kubernetes) and can prevent unauthorized external access, but the clause is general-purpose network services security rather than container-specific, and does not prevent discovery by an actor with valid credentials, from within the cluster, or via non-API vectors such as log leakage.
- T1619partialprevents — A.8.21 can restrict and authenticate access to cloud storage APIs as network services, blocking unauthorized enumeration, but it does not govern object-level list permissions for authenticated or compromised principals, so prevention is partial.
- T1651partialprevents — A.8.21's requirements for authentication, authorization, and monitoring of network services can constrain unauthorized use of cloud administration services, but the technique leverages legitimate admin tools and runs when the adversary holds valid admin or compromised delegated-admin credentials, which the control's requirements do not eliminate.
- T1659partialprevents — A.8.21's encryption, authentication, and network connection controls can prevent content injection against encrypted/authenticated service channels, but the technique's core vectors — upstream ISP-level injection and ManOnTheSide races at network infrastructure — occur outside the organization's service-access controls and remain reachable.
- T1675partialprevents — A.8.21 requires authentication, authorization, and access controls for network services including ESXi management interfaces, which can prevent unauthorized use of the vSphere APIs and VMware Tools Daemon; however, the technique abuses legitimate administration services that must remain accessible, and once an adversary has compromised ESXi/vCenter credentials the controls do not stop the technique from running.
- T1686.002partialprevents — A.8.21 requires authentication, authorization, network management controls, and monitoring for network service access, which if implemented constrain who can modify firewall rules on network devices; however, the control is governance-level and mandates these measures rather than implementing them, and the technique explicitly notes adversaries obtain access via valid accounts or exploitation of vulnerabilities, both of which bypass the requirements A.8.21 sets.
- T1689partialprevents — A.8.21 requires technical parameters for secured connections, authentication/encryption standards, and network connection controls that, when properly specified (e.g., mandating TLS 1.2+ with no fallback, prohibiting HTTP downgrade), prevent the network-protocol variant of T1689; however, the control only addresses the network-services dimension of downgrade attacks and is a requirements-setting governance control whose prevention depends on implementation by network service providers, leaving the PowerShell and boot-manager downgrade vectors uncovered.
Prevented OWASP Web Top 10 (2025) risks (8)
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
- A01partialblocks — A.8.21 explicitly requires formulating and implementing rules plus technological controls for authentication, authorization procedures, network access restrictions, and monitoring, which squarely prevents the authorization-decision slice of A01 (missing checks, incorrect enforcement) while leaving path traversal, CSRF, and IDOR in non-network contexts untouched.
- A02partialblocks — A.8.21 mandates identifying, requiring and monitoring specific security features (authentication, authorization, network connection controls, encryption, access rules) from providers and formulating rules that directly close many misconfiguration vectors in network services, but leaves residual exposure in non-network defaults, application-layer settings, incomplete hardening of unrelated components, and implementation gaps.
- A07partialblocks — A.8.21 explicitly requires authentication requirements, authorization procedures, network connection controls, and monitoring for network services, which directly blocks many authentication bypass, brute-force, credential stuffing, and session hijacking vectors when those services are network-facing; however, it is silent on application-layer specifics such as password reset flows, weak credential storage, or session token implementation details that dominate the OWASP category.
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.