A.8.5 Technological
Secure authentication
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 (25)
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
- IA-2mostlycovers — IA-5 (Authenticator Management) -> A.5.17 is the calibrated anchor for this exact direction and grades mostly; A.8.5's secure-authentication requirements (MFA, credential protections, logon procedures) are substantially but not fully accounted for by IA-2's narrower focus on unique identification plus association with processes.
- IA-5mostlycovers — IA-5's requirements for initial distribution verification, content establishment, strength, and (implied continuation) management are almost entirely accounted for by A.8.5's mandate to ensure secure authentication, but A.8.5 leaves some residual on non-user entities and detailed lifecycle procedures that sit outside its core focus.
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.AA-03fullcovers — The ISO control directly mandates authentication of users, services, and hardware using techniques whose strength matches the sensitivity of the information being accessed.
- PR.AA-01explicit mandateimplements — A.8.5 directly operationalizes secure authentication of identities/credentials, which is the core subject PR.AA-01 requires the organization to manage
- PR.AA-02domain mandateimplements — A.8.5's operational focus on secure authentication directly operationalizes the identity-proofing and credential-binding outcome in PR.AA-02 within the access/identity domain, though the CSF outcome does not name authentication mechanisms specifically
- PR.AA-05explicit mandateimplements — A.8.5 directly operationalizes the enforcement half of PR.AA-05 by providing the secure authentication mechanism that carries out defined and managed access decisions.
- PR.IR-01domain mandateimplements — A.8.5's secure authentication mechanisms give operational effect to protecting networks/environments from unauthorized logical access (PR.IR-01), as authentication is a core technical means within that domain (though not the only one).
Related OWASP ASVS 5.0 requirements (15)
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
- V6.3.1mostlycovers — Guidance on blocking accounts or forcing resets after repeated failed attempts and on raising alerts for suspicious log-on patterns aligns with the ASVS requirement to implement documented controls against brute-force and credential-stuffing attacks.
- V6.3.3mostlycovers — The ISO control's requirement for multi-factor authentication when accessing critical systems directly implements the ASVS mandate that MFA or equivalent combined single-factor mechanisms must be used to reach high-value resources.
Related weaknesses / CWE (2572)
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-293fullblocks — Secure authentication explicitly requires strong, non-spoofable mechanisms, directly preventing reliance on the Referer header.
- CWE-293fullremoves — A.8.5 mandates that a suitable authentication technique be chosen to substantiate the claimed identity of users, software, messages and other entities; using the HTTP Referer field is not a suitable technique, so implementing the clause replaces the Referer-based check with proper authentication and the weakness stops being present, with no remainder to name.
- CWE-306fullremoves — A.8.5 mandates that authentication be in place to substantiate claimed identity when access to systems, applications and services is granted, and CWE-306 is the absence of authentication for critical functions; applying A.8.5 where it was not applied before requires adding authentication, so the weakness stops being present with no remainder, analogous to A.8.24 -> CWE-311 = full.
- CWE-307fullremoves — A.8.5 directly mandates protection against brute force log-on attempts, so implementing it removes the absence of such restrictions that constitutes CWE-307.
- CWE-307fullprevents — A.8.5 item (e) explicitly mandates protection against brute force log-on attempts (CAPTCHA, password reset after N failures, or blocking after a maximum number of errors), which is exactly the measure CWE-307 says is missing; a system designed and implemented under A.8.5 does not acquire this weakness, with no nameable remainder.
- CWE-308fullremoves — A.8.5 directly mandates multi-factor authentication for critical information systems and requires authentication strength appropriate to information classification, which is exactly the remedy for CWE-308 (single-factor where more is required). When applied to a system that previously used single-factor authentication in a context requiring MFA, the weakness is removed — this is the 'mandates a measure removes the weakness that is the absence of that measure' pattern, and the clause's provisions cover the full scope of where CWE-308 applies.
- CWE-523fullremoves — A.8.5 guidance item (j) directly mandates 'not transmitting passwords in clear text over a network', which is precisely the measure whose absence is CWE-523; applied to a system that previously transmitted credentials unprotected, implementing the clause requires enabling transport protection (e.g., TLS on the login page), so the existing weakness stops being present with no remainder.
- CWE-290mostlyremoves — A.8.5 mandates strong authentication methods (certificates, tokens, biometrics) and MFA for critical systems, so applying it to a system that relied on a single spoofable factor (IP address, DNS name) replaces that factor with non-spoofable ones and the spoofing attack no longer succeeds; the remainder is implementation flaws inside otherwise strong authentication (e.g. flawed certificate validation), which the clause's guidance on choosing techniques does not reach.
- CWE-291mostlyremoves — A.8.5 mandates suitable authentication techniques and MFA for critical systems, so applying it where IP-only authentication existed replaces that sole-factor reliance with proper authentication, removing the CWE-291 weakness in the bulk of cases; the remainder is bounded — non-critical subsystems where A.8.5's MFA mandate does not bite and where 'suitable technique' is read loosely enough to leave an IP check in place as the primary mechanism.
- CWE-291mostlyprevents — A.8.5's guidance on choosing suitable authentication techniques, with strength appropriate to information classification, and its MFA requirement for critical systems, prevents a well-implemented organization from designing systems that rely solely on IP address for authentication; the bounded remainder is low-value resources where 'appropriate strength' might permit weak single-factor schemes, plus implementations that fall short of the clause's intent.
- CWE-300mostlyblocks — Secure authentication mechanisms verify actor identity at both ends of the channel.
- CWE-308mostlyprevents — A.8.5 directly mandates multi-factor authentication for critical information systems and prescribes authentication methods alternative to passwords where strong authentication is required, which is precisely the design choice that creates CWE-308; a well-implemented control prevents the weakness in the contexts where it applies, with the remainder being organizational judgments about which systems qualify as 'critical' or warrant stronger authentication.
- CWE-309mostlyremoves — A.8.5 mandates multi-factor authentication for critical information systems, which removes the password-as-primary weakness for those systems; the remainder is bounded and nameable — non-critical systems where A.8.5 does not mandate MFA, so passwords may still be the sole authentication factor there.
- CWE-322mostlyblocks — Secure authentication directly prevents unauthenticated key exchange by requiring verified identities before keys are established.
- CWE-603mostlyblocks — Requires server-side authentication mechanisms that directly prevent client-only checks.
- CWE-613mostlyremoves — A.8.5 items (k) and (l) directly mandate session termination after inactivity and connection duration restrictions, which are the primary mechanisms for session expiration and address the core of CWE-613; applying the clause to a system that lacked these removes the timeout-based weakness, but the CWE has been expanded to include logout invalidation and session fixation variations that A.8.5 does not reach, and non-configurable third-party products (the majority of observed CVE instances) cannot be remediated by this clause alone.
- CWE-613mostlyprevents — A.8.5 items (k) and (l) directly mandate session termination after inactivity and connection duration limits, which are the primary mechanisms for preventing insufficient session expiration; the nameable remainder is logout invalidation, a variation folded into CWE-613 that the clause does not explicitly address.
- CWE-613mostlyblocks — A.8.5 items (k) and (l) directly mandate terminating inactive sessions after a defined period and restricting connection duration times, which blocks the core exploitation of CWE-613 (reusing old session credentials indefinitely); the bounded remainder is the variations the clause does not reach — sessions not invalidated on logout and session fixation — which are nameable and outside the authentication-procedure scope.
- CWE-654mostlyremoves — A.8.5 mandates multi-factor authentication for critical information systems and requires authentication methods alternative to passwords where strong authentication is required; applied to a system that previously used single-factor authentication, it requires adding factors, which removes the CWE-654 weakness for those systems. The remainder is non-critical, low-risk systems where the clause only requires 'appropriate' strength and single-factor remains acceptable by design.
- CWE-654mostlyprevents — A.8.5 explicitly mandates multi-factor authentication for critical information systems and requires authentication strength appropriate to information classification, so a system under it does not acquire single-factor reliance for critical access — the weakness is kept out at design. The bounded remainder is non-critical systems where the clause's own guidance permits single-factor authentication as appropriate to the classification.
- CWE-1254partialblocks — A.8.5's brute-force protection (item e) limits authentication attempts, which makes the multi-attempt timing attack on non-constant-time comparisons impractical in many deployments; however, the weakness itself (the non-constant-time comparison) remains present and exploitable in scenarios without strict lockout or where many accounts are available, and the clause does not address the comparison-logic implementation at all.
- CWE-1391partialblocks — A.8.5 blocks exploitation of weak user credentials through MFA (required for critical systems) and brute-force protection (item e: lockout, CAPTCHA), but does not block exploitation of hard-coded service credentials, default credentials that are publicly known, or previously-leaked credentials where the attacker already possesses the secret and brute-force resistance is irrelevant.
- CWE-1392partialblocks — A.8.5 mandates multi-factor authentication for critical information systems, which blocks exploitation of default credentials where MFA is in place (the default password alone is insufficient without the additional factor); however, MFA is only required for critical systems, and A.8.5 does not address default credentials directly, so non-critical systems and instances where MFA is not deployed remain exploitable.
- CWE-1393partialblocks — A well-implemented A.8.5 blocks exploitation of default passwords where it imposes MFA or alternative authentication for critical or high-assurance access, but it does not itself require changing default credentials, so password-only systems remain exploitable.
- CWE-1428partialremoves — A.8.5 item (j) requires that passwords not be transmitted in cleartext over a network, which removes the weakness in the authentication flow (typically forcing HTTPS for login), but the clause does not address the product's broader reliance on HTTP for non-authentication communications, leaving the majority of the weakness intact.
- CWE-1428partialprevents — A.8.5(j) requires passwords not be transmitted in clear text over a network, which prevents the narrow slice of CWE-1428 where authentication credentials traverse HTTP, but the broader weakness — any product reliance on HTTP for application data, APIs, or other traffic — is outside the clause's scope.
- CWE-200partialblocks — Suppressing system details, error specifics, and previous log-on information until successful authentication reduces the information an unauthenticated attacker can gather.
- CWE-201partialprevents — A.8.5 prevents sensitive authentication-related data such as passwords and pre-login system details from being inserted into logon traffic, but its scope is authentication and it does not reach the broader class of sensitive information embedded in other application responses or error messages.
- CWE-203partialblocks — A.8.5's log-on procedure guidance (items c, d, e) directly blocks user enumeration via error messages and timing discrepancies during authentication, but the control does not address observable discrepancies outside the log-on process such as side-channel attacks or behavioral information leaks.
- CWE-203partialremoves — A.8.5's log-on procedure guidance removes existing authentication-flow observable discrepancies such as user enumeration and differing error messages, but it does not reach non-authentication side-channel or timing discrepancies in the broader CWE-203 class.
- CWE-204partialblocks — Strong authentication reduces information leakage from authentication responses.
- CWE-205partialremoves — A.8.5's log-on procedure guidance items (a, c, i, j) directly address authentication-related behavioral discrepancies — not displaying sensitive info before logon, not providing help messages that aid unauthorized users, not displaying passwords in cleartext, not transmitting passwords in cleartext — which are instances of CWE-205. When applied to a system that previously lacked these controls, those specific flaws are removed. However, CWE-205 is a broad class covering network protocol behaviors (TCP/IP stack, ICMP, SYN-FIN), application fingerprinting, and system footprinting, the vast majority of which fall outside A.8.5's authentication scope.
- CWE-205partialprevents — A.8.5 guidance items (a), (c), (i), (j) directly prevent observable behavioral discrepancies in the log-on procedure (e.g., not indicating which part of credentials is wrong, not displaying passwords in cleartext, not transmitting passwords in cleartext), but CWE-205 extends well beyond authentication to network stack behavior, application fingerprinting, and other product-level behavioral leaks that A.8.5 does not reach.
- CWE-206partialremoves — A.8.5 guidance item (c) directly mandates uniform log-on error messages that do not indicate which part of the data is correct or incorrect, which is the canonical authentication-context instance of CWE-206 (username enumeration via differential error responses); implementing the clause where it was not applied before removes that subset of the weakness, but CWE-206 also arises in non-authentication contexts (file existence infoleaks, group enumeration, timing discrepancies in other product behaviors) that A.8.5 does not reach, leaving a nameable remainder.
- CWE-206partialprevents — A.8.5 item (c) directly prescribes against the canonical login-procedure form of CWE-206 by requiring that error conditions not indicate which part of the data is correct or incorrect, so a well-implemented logon procedure does not exhibit the username-enumeration / behavioral-discrepancy weakness at all; the remainder is bounded and nameable — observable behavioral discrepancies arising in non-login contexts (file-existence infoleaks, group-name enumeration, GUI timing leaks) are outside the clause's scope.
- CWE-208partialblocks — A.8.5 items (c) and (d) directly address timing-based information disclosure during log-on — not indicating which part of the data is correct/incorrect, and validating log-on information only on completion of all input — which blocks the byte-by-byte password comparison and differential-error timing attacks documented in CWE-208's observed examples; however, the clause does not reach timing discrepancies in cryptographic operations (HMAC comparison, padding checks) or other authentication-related timing side channels beyond the log-on procedure itself.
- CWE-209partialremoves — A.8.5 guidance items (a) and (c) specifically address error messages during the log-on procedure — not displaying sensitive information until log-on completes, and not providing help messages that would aid an unauthorized user (e.g., not indicating which part of the data is correct or incorrect). This removes a narrow subset of CWE-209 (log-on error message leaks), but the vast majority of CWE-209 instances — database errors, file processing errors, configuration errors, and general application error handling — are outside A.8.5's scope.
- CWE-209partialprevents — A.8.5 item (c) directly prevents a specific subset of CWE-209 — log-on error messages that would aid an unauthorized user by indicating which part of the data is correct or incorrect — but the vast majority of CWE-209 instances occur in non-authentication contexts (database errors, file processing, business logic) that this clause does not reach.
- CWE-210partialprevents — A.8.5 guidance item (c) explicitly prevents log-on error messages from indicating which part of credentials is correct/incorrect, addressing a narrow slice of CWE-210; the vast majority of self-generated error messages containing sensitive information (business logic, file system, database errors elsewhere in the application) are outside the authentication clause's scope.
- CWE-210partialremoves — A.8.5 guidance item (c) directly mandates that log-on error messages not reveal which part of the data is correct or incorrect, so when applied to a system that previously leaked such information during authentication, those specific instances of CWE-210 are removed; however, CWE-210 covers self-generated error messages containing sensitive information throughout the application (paths, configuration details, internal state), and A.8.5's reach is confined to the log-on procedure, leaving the broader class untouched.
- CWE-223partialremoves — A.8.5 mandates logging of successful and unsuccessful authentication attempts (item f), raising security events on potential breaches (item g), and displaying previous logon details (item h), which directly removes the authentication-related instances of CWE-223 (e.g., CVE-1999-1029, CVE-2000-0542); however, CWE-223 is broader than authentication logging and includes omissions in other contexts (e.g., CVE-2002-1839 email sender IP, file access canonicalization) that A.8.5 does not reach.
- CWE-223partialprevents — A.8.5 mandates logging of unsuccessful and successful authentication attempts, raising security events on potential breaches, and displaying prior logon details — which prevents the authentication-related subset of CWE-223 (e.g., the observed examples about login attempts not being recorded) from arising. The remainder is the broader class: omission of security-relevant information outside the authentication context (sender IPs in email, file access details, etc.), which this clause does not reach.
- CWE-223partialblocks — A.8.5 mandates logging of unsuccessful and successful authentication attempts (item f), raising security events on potential breaches (item g), and displaying prior logon information (item h), which blocks exploitation of CWE-223 in the authentication context — attackers cannot hide their login attempts because the events are recorded. However, CWE-223 extends well beyond authentication (e.g., CVE-2002-1839 about sender IP not recorded in outgoing email, file-access logging with canonical names), and A.8.5 does not reach those contexts.
- CWE-262partialblocks — A.8.5 mandates MFA for critical systems and recommends strong non-password authentication (certificates, tokens, biometrics), both of which block exploitation of an unaged password — but the weakness remains fully exploitable on non-critical systems using single-factor password auth, which the clause does not reach.
- CWE-287partialblocks — A.8.5 mandates secure authentication techniques, multi-factor authentication, and log-on procedures that directly prevent unauthorized access, fully blocking exploitation of improper authentication weaknesses.
- CWE-289partialblocks — Secure-authentication controls can require canonical identifiers and reject alternate names.
- CWE-290partialprevents — A.8.5 mandates strong authentication methods (digital certificates, tokens, biometrics) and multi-factor authentication for critical systems, which directly addresses CWE-290's root cause of relying on spoofable identifiers like IP addresses or DNS names; the remainder is bounded and nameable—spoofing that survives proper MFA (e.g., session token theft, certificate authority compromise in complex PKI)—which is what keeps this from `full`.
- CWE-293partialprevents — A.8.5's purpose covers authentication of 'messages and other entities' and mandates choosing suitable authentication techniques, which would steer a well-implemented policy away from using a spoofable HTTP header as authentication; however, the clause's specific guidance items (a–l) are all about user log-on procedures, so the design pattern of referer-based request validation is only reached by the general principle, not by the operational guidance.
- CWE-294partialblocks — A.8.5 addresses several vectors relevant to capture-replay: item (j) bars cleartext password transmission, items (k) and (l) shorten the window via session termination and connection-duration limits, and the strong-authentication and MFA guidance (certificates, smart cards, tokens) typically carries replay protection as part of those mechanisms. But the clause does not mandate anti-replay primitives (nonces, timestamps, challenge-response) at the protocol level, and for non-critical systems password-only authentication remains acceptable, so a system compliant with A.8.5 can still be vulnerable to capture-replay — the remainder is nameable and not bounded to a vanishing slice.
- CWE-294partialremoves — A.8.5 mandates practices that remove some capture-replay vectors when applied to a system that previously lacked them: item (j) prevents cleartext password transmission (the most common capture vector), MFA requirements mean a captured factor alone cannot replay authentication, and session termination/duration limits (k, l) shrink the replay window. However, the clause governs authentication procedures and password handling, not the cryptographic design of authentication protocols generally; capture-replay of session tokens, authentication challenges, or other non-password messages is not squarely addressed, leaving a nameable remainder of the weakness class.
- CWE-294partialprevents — A.8.5 mandates MFA for critical systems and explicitly prohibits transmitting passwords in cleartext over a network (item j), which together prevent most capture-replay: the captured credential alone is insufficient under MFA, and cleartext password transmission — the primary capture vector — is forbidden. The remainder is bounded and nameable: session-token replay after MFA authentication, and capture-replay against non-critical systems where MFA is not mandated, both fall outside the clause's authentication-stage scope.
- CWE-302partialblocks — Secure authentication explicitly requires server-side validation of all identity claims, directly eliminating this weakness.
- CWE-303partialblocks — Secure authentication control directly requires correct implementation of authentication algorithms.
- CWE-305partialblocks — Secure authentication mechanisms directly address bypass risks in the implemented authentication process.
- CWE-307partialblocks — CAPTCHA, account lock-out after repeated failures, and alerts on excessive attempts directly stop automated brute-force guessing of credentials.
- CWE-309partialblocks — A.8.5 blocks some exploitation paths of password-based authentication through brute-force protection (item e), cleartext transmission prevention (item j), logging and alerting (items f, g), and session controls (items k, l), but the primary mechanism against CWE-309 is prevention via MFA and stronger-method requirements, and the blocking mechanisms leave credential stuffing, phishing, keylogging, and password reuse attacks unaddressed.
- CWE-309partialprevents — A.8.5 mandates multi-factor authentication for critical systems and alternative authentication methods where strong authentication is required, directly preventing password-only authentication in those cases; the bounded remainder is non-critical or lower-classification systems where password-only remains acceptable under the clause.
- CWE-319partialremoves — A.8.5 requires that passwords not be transmitted in cleartext, so applying it removes that specific instance of cleartext transmission, but it does not address other sensitive information sent in cleartext.
- CWE-319partialprevents — A.8.5 item (j) directly prevents cleartext transmission of passwords over a network, and item (i) prevents cleartext display during entry, addressing a meaningful subset of CWE-319 (authentication credentials). However, CWE-319 covers all sensitive or security-critical data in cleartext — session tokens, encryption keys, configuration data, PII, financial data — which A.8.5 does not reach, so only a minority of instances are prevented.
- CWE-322partialremoves — A.8.5 mandates authentication of users and entities, which directly addresses CWE-322's absence of entity authentication in key exchange; however, the clause's specific guidance is about log-on procedures (passwords, MFA, brute-force protection, session termination), not about cryptographic protocol design, so it removes the weakness only where the implementation extends to key-exchange entity authentication, leaving protocol-level cases uncovered.
- CWE-334partialblocks — A.8.5(e) and (g) mandate brute-force protections (CAPTCHA, lockout after failed attempts, security event on threshold) that directly counter the exploitation method for CWE-334's small random value space at log-on, but the weakness itself remains in the product and exploitation via other vectors (predictable generation, non-authentication uses like session IDs) is outside the clause's reach.
- CWE-350partialprevents — A.8.5 mandates that a 'suitable authentication technique' be chosen and that authentication strength be appropriate to the information's classification, which would steer a well-implemented organization away from using spoofable reverse DNS as a security-critical authentication factor; however, the clause addresses authentication methods generally (MFA, log-on procedures, biometrics) and does not specifically prohibit reverse-DNS-based authentication, so the effect depends on the implementer's reading of 'suitable' and leaves a nameable remainder of instances where reverse DNS is still chosen.
- CWE-358partialblocks — Secure authentication control may be undermined if the underlying protocol checks are improperly implemented.
- CWE-408partialblocks — Secure authentication ensures identity verification occurs before expensive operations are allowed.
- CWE-497partialprevents — A.8.5 items (a) and (c) directly prevent sensitive system information and help messages from being displayed during the log-on procedure, which is a real slice of CWE-497; the remainder — system information exposed through error handling, debugging output, diagnostic messages, and logging elsewhere in the product — is outside the clause's scope.
- CWE-5partialprevents — A.8.5 directly prevents the credential-transmission slice of CWE-5, such as clear-text passwords over the network, but leaves the broader unencrypted application data and J2EE SSL misconfiguration largely unaddressed.
- CWE-521partialprevents — A.8.5 requires authentication strength appropriate to information classification and mandates MFA for critical systems, which implies password strength requirements when passwords are used, but the clause text focuses on authentication mechanisms, MFA, and log-on procedures rather than explicitly mandating password complexity, length, or composition rules — so it addresses the weakness indirectly and a well-implemented version would include password policies, but the remainder (specific password strength requirements) is not directly mandated by the clause.
- CWE-521partialblocks — A.8.5's log-on procedure guidance (e) provides brute-force protection via lockouts and attempt limits, and (g) raises alerts on suspicious attempts, which block some exploitation of weak passwords; however, the weakness (allowing weak passwords) remains present and is still exploitable via offline cracking, credential stuffing, social engineering, and limited dictionary attacks within the lockout threshold.
- CWE-523partialblocks — A.8.5 mandates not transmitting passwords in clear text over a network (item j), which directly blocks exploitation of CWE-523 for the password portion of credentials, but the clause does not mandate encrypted transport for the username or for other authentication tokens, leaving a nameable remainder of the weakness still exploitable.
- CWE-537partialprevents — A.8.5's log-on procedure rules keep some authentication-flow error messages from exposing sensitive details, but the clause does not govern general Java runtime exception handling where most CWE-537 instances arise.
- CWE-550partialprevents — A.8.5 guidance item (a) prevents sensitive information from being displayed during the log-on process, addressing a narrow slice of CWE-550; the vast majority of server-generated error message weaknesses occur outside the authentication context (database errors, network failures, application logic) where A.8.5 has no reach.
- CWE-620partialblocks — Mandates secure authentication mechanisms, which include verifying existing credentials before permitting password changes.
- CWE-693partialremoves — A.8.5 mandates authentication mechanisms (MFA, strong authentication, secure log-on procedures), so applied to a system that previously lacked them it removes the authentication-related instances of CWE-693; but CWE-693 is a pillar covering protection mechanism failures across encryption, access control, sandboxing, input validation and many other domains, of which authentication is only one slice, so the remainder is the broad majority of the class.
- CWE-693partialprevents — A.8.5 mandates authentication protection mechanisms (MFA, brute-force protection, secure log-on procedures), so authentication-related instances of CWE-693 do not arise under it; but CWE-693 is a pillar weakness covering protection mechanisms of all kinds (ACLs, sandboxing, encryption, virtualization, session management, etc.), and A.8.5 reaches only the authentication slice, leaving the broader class untouched.
- CWE-778partialblocks — A.8.5 mandates logging of unsuccessful and successful log-on attempts (item f), which directly blocks the exploitability of CWE-778's insufficient logging for authentication events, but the control's scope is limited to log-on procedures and does not address logging gaps for other security-critical events.
- CWE-778partialprevents — A.8.5 items (f) and (g) mandate logging of successful and unsuccessful authentication attempts and raising security events on potential breaches, so the authentication-event subset of CWE-778 is prevented from arising in a well-implemented system; the remainder is the broader weakness — logging of non-authentication security-critical events — which A.8.5 does not reach.
- CWE-778partialremoves — A.8.5 requires logging of successful and unsuccessful authentication attempts, so retrofitting it can close existing logging gaps for authentication events, but CWE-778 covers security-critical events generally and non-authentication logging gaps remain untouched.
- CWE-804partialblocks — Strong authentication mechanisms include robust CAPTCHA implementations that resist automated guessing.
Mitigated MITRE ATT&CK techniques (3019)
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 — A.8.5's MFA and strong authentication requirements prevent the adversary's stolen credentials from successfully authenticating to remote services like SSH/RDP/VNC, mirroring the T1110.001 anchor; the same bounded remainder applies — identities exempted from MFA, legacy protocols that bypass it, and service accounts on static secrets remain reachable for lateral movement.
- T1021.005mostlyprevents — A.8.5's requirements for strong authentication, MFA on critical systems, and brute-force protections (item e: CAPTCHA, password reset, account lockout) directly address VNC's reliance on valid-account authentication and its known susceptibility to brute-force attacks; the bounded remainder is VNC configured with its own authentication outside the system's MFA scope, VNC implementations with memory-corruption vulnerabilities that bypass authentication entirely, and exempted service accounts — the same shape the T1110.001 anchor names.
- T1021.007mostlyprevents — A.8.5's MFA and strong authentication requirements prevent most uses of stolen or guessed credentials against cloud services, but the technique's federation pathway (where on-premises authentication compromise is trusted by the cloud) and Application Access Tokens bypass username/password controls, leaving a named remainder analogous to the T1110.001 anchor.
- T1055.003mostlydetects — A.8.5 requires logging of successful/unsuccessful log-on attempts, raising security events/alerts on potential breaches or repeated failures, and monitoring for anomalous access patterns (unusual location/device/time) that can surface thread-injection anomalies when they trigger authentication or session events; this covers most observable cases but leaves a bounded remainder for fully in-memory hijacks with no logon or anomaly trigger.
- T1078mostlyprevents — Mandating MFA and non-password authenticators raises the bar for attackers who have obtained valid credentials but still need an additional factor.
- T1078.001mostlyprevents — A.8.5's MFA and strong-authentication requirements can prevent abuse of default accounts when properly applied to them, but the control does not address the root cause (the existence of default accounts, which is A.8.9's territory), default accounts often bypass MFA policies (service accounts, built-in OS accounts), and the technique includes credential theft via exploitation and stolen private keys that A.8.5 does not reach.
- T1078.002mostlyprevents — A.8.5's MFA and strong-authentication requirements block most uses of stolen domain credentials, but the same named remainder as the T1110.001 anchor applies: identities exempted from MFA, legacy protocols that bypass it, and service accounts on static secrets stay reachable to valid-credential abuse.
- T1078.004mostlyprevents — A.8.5 mandates multi-factor authentication and strong authentication techniques appropriate to the information's classification, which prevents the use of compromised or guessed credentials for cloud accounts where enforced, but the technique remains reachable through misconfigurations, legacy federation, or accounts exempted from MFA.
- T1110mostlyprevents — Locking accounts or requiring resets after repeated failed attempts, plus CAPTCHA, directly limits the number of guesses an attacker can make before being blocked.
- T1110mostlydetects — A.8.5 explicitly mandates logging of unsuccessful and successful attempts (f), raising a security event on potential brute-force breach (g), and alerting administrators when wrong-password thresholds are reached — direct detection of online brute force. The remainder is offline brute force against previously stolen hashes, which the authentication log-on procedure does not see.
- T1110.001mostlyprevents — The same brute-force protections and account lockouts reduce the feasibility of automated password guessing against individual accounts.
- T1110.001mostlydetects — A.8.5 explicitly requires logging of unsuccessful/successful log-on attempts, raising security events/alerts on potential breaches (e.g. after failed attempts), and displaying prior failed log-ons, which surfaces password guessing in flight or after the fact across the technique's core login vectors.
- T1110.001mostlyresponds — A.8.5 explicitly requires raising a security event and sending alerts on detected breach of log-on controls (including brute-force guessing attempts), which is the core act of `responds` once the technique is underway; the remainder is that the clause does not itself perform containment or eradication.
- T1110.002mostlydetects — A.8.5 explicitly requires logging of successful/unsuccessful log-on attempts, raising security events/alerts on potential breaches (e.g. repeated failures), and other observable indicators that surface password-cracking attempts in flight or post-facto on the monitored estate.
- T1110.003mostlyprevents — Rate-limiting and lockout mechanisms blunt password-spraying campaigns that rely on many rapid, low-and-slow attempts across numerous accounts.
- T1110.004mostlyprevents — A.8.5's MFA requirement and brute-force protections (item e: lockout after failed attempts, CAPTCHA) directly defeat credential stuffing, and the technique's own description names login-failure policies as the primary counter; the same remainder as the T1110.001 anchor applies — identities exempted from MFA, legacy protocols that bypass it, and service accounts on static secrets stay reachable, and slow-and-low attacks can evade lockout thresholds.
- T1550.002mostlyprevents — A.8.5's MFA requirement defeats PtH where it is enforced, since possessing a hash alone is insufficient without the second factor; the bounded remainder is the same shape as the T1110.001 anchor -- identities exempted from MFA, legacy NTLM-only systems, and service accounts on static secrets remain reachable to hash-based authentication.
- T1586mostlyprevents — A.8.5 mandates strong/multi-factor authentication, brute-force protections, MFA for critical systems, session termination, and no cleartext passwords/transmission, which directly blocks the credential-compromise vectors (phishing, brute force, reuse from dumps) named in T1586; the bounded remainder is non-technical vectors such as paying insiders.
- T1586.002mostlyprevents — A.8.5 mandates MFA, brute-force protections, no cleartext passwords, session timeouts and strong auth proportionate to data sensitivity, which directly stops the credential-theft vectors (phishing, brute force, password reuse) that enable T1586.002; the bounded remainder is non-technical vectors such as insider sales of credentials or post-compromise thread hijacking.
- T1586.003mostlyprevents — A.8.5 directly counters the dominant credential-compromise vectors named for T1586.003 (password spraying, phishing-for-credentials, brute-force, clear-text transmission, weak log-on procedures) by mandating strength-proportionate MFA, anti-brute-force measures, no clear-text passwords, secure log-on design, and alerts; the bounded remainder is purchase of already-breached credentials and theft of application access tokens that bypass authentication entirely.
- T1589mostlyprevents — A.8.5 mandates MFA, strong auth, brute-force protections, no cleartext passwords, session timeouts and log-on hardening that directly stop many forms of credential/MFA-configuration harvesting (phishing, username enumeration, sniffing, brute-force) from succeeding; the named remainder is passive OSINT from public leaks or social media that the control does not address.
- T1589.001mostlyprevents — A.8.5 mandates MFA, strong auth, brute-force protections, no cleartext passwords, session timeouts and log-on hardening that stop many credential-gathering vectors (phishing elicitation, cookie theft, brute-force, sniffing, reuse) from succeeding; the named remainder is credentials already exposed via breaches, dark-web markets or infostealer logs that the control does not block from being gathered.
- T1608.005mostlyprevents — A.8.5 mandates strong/MFA authentication, anti-brute-force measures, secure log-on procedures, and no clear-text credential transmission; these directly stop credential-harvesting link targets (phishing pages) from succeeding when the user reaches them, though they do not stop the adversary from staging the link target itself.
- T1021.001partialprevents — A.8.5 requires MFA for critical systems and protects against brute-force log-on attempts, which can prevent RDP access via guessed or stolen credentials, but it does not remove the RDP service or prevent access using valid credentials obtained through other means.
- T1021.002partialprevents — A.8.5's MFA and log-on protections (e.g., failed-attempt limits) can block unauthorized SMB access via stolen credentials, but the technique also works with valid accounts that pass authentication, leaving a reachable remainder.
- T1021.004partialdetects — A.8.5's clauses (f) logging unsuccessful and successful attempts and (g) raising a security event on potential breach of log-on controls surface SSH authentication events, but detection stops at the log-on boundary and does not cover the post-authentication remote shell activity that is the bulk of T1021.004.
- T1021.004partialprevents — A.8.5 prevents some instances of T1021.004 where MFA is enforced on SSH access and where the valid account relies on a password that can be blocked by brute-force protections, but the technique explicitly supports public-private keypair authentication which can bypass password-based MFA, and exempted/service accounts and key-based-only configurations remain reachable, leaving a substantial remainder.
- T1021.005partialdetects — A.8.5's logging of successful/unsuccessful attempts (f) and security event raising on potential breaches (g) detect VNC authentication events when VNC uses system credentials, but the control does not detect VNC sessions using VNC-specific credentials, already-authenticated sessions, or the VNC protocol itself — only the authentication surface it governs.
- T1021.006partialprevents — A.8.5 requires MFA for critical systems and strong authentication proportionate to classification, which can block WinRM access when enforced, but the control sets a requirement rather than a universal mechanism, leaving reachable gaps where MFA is not mandated, legacy protocols bypass it, or service accounts rely on static secrets.
- T1021.008partialprevents — A.8.5's MFA and strong-authentication requirements apply to cloud VM connection methods (Azure Serial Console, EC2 Instance Connect, SSM); where MFA is enforced and the adversary holds only a password, the technique cannot authenticate and is prevented. The named remainder is the case where the adversary possesses the full credential set — a stolen SSH key, a valid application access token, or an identity exempted from MFA on these specific cloud-native paths — which is the same shape of bounded remainder the T1110.001 anchor carries.
- T1040partialprevents — A.8.5 clause (j) directly requires that passwords not be transmitted in clear text over a network, which prevents the credential-capture subset of T1040, but the clause does not address sniffing of configuration details, traffic volumes, headers, or other non-authentication data that the technique also harvests.
- T1072partialprevents — A.8.5 with MFA and strong authentication prevents the credential-stuffing/guessing path to the deployment tool, but T1072 also runs via already-authenticated sessions, service/machine accounts that bypass interactive MFA, local agent abuse, and exploitation of vulnerabilities in the tool itself — none of which A.8.5 reaches.
- T1078.002partialdetects — A.8.5 includes logging of successful and unsuccessful attempts (f) and raising security events on potential breaches (g), which provides some detection surface, but for T1078.002 the abuse uses valid domain credentials so authentication succeeds legitimately and the control's brute-force-oriented detection does not flag the misuse; what is detected is the authentication event itself, not the adversarial abuse.
- T1078.003partialprevents — A.8.5 requires MFA and strong authentication for critical systems, which can prevent abuse of local accounts on those systems, but local accounts on non-critical systems or those exempted from MFA remain reachable.
- T1078.004partialdetects — A.8.5 includes detection mechanisms — logging successful and unsuccessful attempts, raising security events on potential breaches, notifying users of previous logon times and failed attempts, and triggering additional factors on unusual location/device/time patterns — that surface use of valid cloud accounts. These are secondary to the control's preventive purpose and do not detect credential compromise itself or all uses of legitimately-held credentials, so the coverage is a slice rather than the bulk.
- T1098.005partialprevents — A.8.5 mandates MFA and strong authentication, which is the control surface this technique targets, but the technique explicitly exploits MFA self-enrollment flows that require only username/password to register a new device — a gap A.8.5 does not specifically address, so where enrollment bypasses MFA the control does not prevent the registration.
- T1110.002partialprevents — A.8.5's requirements for authentication strength appropriate to classification and MFA reduce the value of cracked credentials and make hash cracking harder/longer, but the offline cracking itself happens on adversary infrastructure outside the control's scope, and rate-limiting/lockout protections (item e) do not apply to offline attacks, so the technique can still run and succeed against weaker secrets.
- T1110.003partialdetects — A.8.5 requires logging of unsuccessful and successful attempts and raising a security event when a threshold of wrong password attempts is reached, which surfaces password spraying in flight, but the control's primary mechanism is authentication hardening rather than monitoring depth, so detection is a byproduct and not the control's full reach.
- T1110.004partialdetects — A.8.5 requires logging of unsuccessful and successful attempts and raising a security event after a threshold of failures, which surfaces credential stuffing in flight, but the control's primary mechanism is prevention and detection depends on the logging and alerting configuration being implemented and monitored.
- T1114.002partialprevents — A.8.5's multi-factor authentication and log-on procedure hardening can prevent unauthorized access to email servers using stolen credentials, but the control does not address access via compromised access tokens or sessions that bypass the authentication step.
- T1133partialdetects — A.8.5's logging of successful/unsuccessful attempts and raising of security events on potential log-on breaches detects authentication-based external remote service access (VPN/Citrix/RDP with valid or guessed credentials), but does not detect the non-authenticated paths in T1133 — exposed Docker/Kubernetes APIs, services without authentication, or Tor hidden services — which leave no log-on event for the clause to surface.
- T1133partialprevents — A.8.5 prevents the credential-based vector of T1133 (MFA, strong auth, brute-force protection stop access via stolen or guessed credentials to VPNs/Citrix/etc.), but the technique also includes unauthenticated exposed services (Docker/Kubernetes APIs, web dashboards) and Tor hidden service persistence, which A.8.5 does not address.
- T1199partialprevents — A.8.5's authentication requirements (especially MFA) raise the bar for an adversary using compromised third-party credentials, but T1199 fundamentally exploits the existing trusted access grant itself — which is outside A.8.5's scope — and even strong authentication will still authenticate the (now-compromised) third party, so the technique is constrained, not prevented.
- T1496.003partialdetects — A.8.5's authentication logging (clause f) and security-event raising (clause g) capture OTP/verification requests, so a flood of them — the mechanism SMS pumping uses — is visible in the logs the clause mandates; but the control is designed for authentication integrity, not for detecting messaging-infrastructure abuse, so it surfaces a slice (request volume anomalies) rather than the financial/resource impact that defines the technique.
- T1530partialprevents — A.8.5 prevents the credential-abuse vector of T1530 (MFA and strong authentication block access via leaked or stolen cloud credentials, and brute-force protections defeat guessing), but it does not address the misconfiguration vector that the technique description highlights as the common case — public buckets, overly-broad IAM, and anonymous access bypass authentication entirely, so a substantial share of T1530 incidents remain reachable.
- T1538partialprevents — A.8.5 requires MFA and strong authentication, which can prevent stolen credentials from granting dashboard access, but the control sets a requirement rather than a universal mechanism, so identities exempted from MFA or using legacy protocols remain reachable.
- T1550partialprevents — A.8.5's MFA and strong-authentication requirements (certificates, smart cards, tokens) can prevent some uses of alternate authentication material where a stolen hash or ticket alone is insufficient, but the control addresses the authentication process rather than credential storage or token protection, so stolen material that the system has already validated (Kerberos tickets, session cookies, application tokens) remains accepted and the technique still runs in many cases.
- T1550.004partialprevents — A.8.5's session termination after inactivity (k) and connection duration limits (l) bound the validity window of any session, including a stolen cookie, partially preventing the technique; however, the technique explicitly bypasses MFA and strong authentication — A.8.5's primary mechanisms — because the cookie represents an already-authenticated session, so the core of the technique remains executable within the session window.
- T1552.007partialprevents — A.8.5's strong authentication and MFA requirements prevent unauthorized access to Docker/Kubernetes APIs, but the technique explicitly covers adversaries with sufficient permissions (e.g., via a pod's service account) who are already authenticated, and credential exposure in API logs is a storage/logging concern outside A.8.5's scope.
- T1555.005partialprevents — A.8.5's brute-force protection (item e) and MFA requirements partially address the password-guessing vector of T1555.005 (which itself references T1110.001 for brute-forcing the master password), but the technique's main vectors — extracting credentials from memory after the database is unlocked and exploiting password-manager vulnerabilities — are outside the control's scope, which is system authentication rather than securing third-party credential stores.
- T1556.001partialprevents — A.8.5's MFA requirement limits the technique's effectiveness — Skeleton Key provides 'unfettered access' only in single-factor environments, so MFA blocks the bypass from succeeding for protected accounts — but the control does not prevent the patching of the domain controller's authentication process itself, and accounts or factors outside MFA remain reachable.
- T1557.001partialprevents — A.8.5's MFA requirement prevents the relay component (a captured NTLMv2 hash alone cannot satisfy a second factor), and its strength-of-authentication guidance makes offline cracking harder, but the clause does not address the root cause — insecure name resolution protocols — and hash capture still occurs; MFA is conditional on system classification rather than universal, and the poisoning itself is outside the clause's scope.
- T1557.004partialprevents — A.8.5's MFA requirement prevents captured credentials from being reused against the legitimate service, and certificate-based authentication would resist fake captive portals, but the technique's network-level effects (sniffing, data manipulation) and the connection to the rogue AP itself sit outside what authentication controls address.
- T1558.004partialprevents — A.8.5's MFA requirement for critical systems and strong-authentication mandate would block AS-REP Roasting where applied, but the clause does not address the core enabler (Kerberos preauthentication disabled on the account) or weak encryption algorithms (e.g., RC4), and MFA is scoped to critical systems only, leaving a substantial remainder.
- T1563partialprevents — A.8.5's session-management clauses (k: terminating inactive sessions, l: restricting connection duration) reduce the window in which a hijackable session exists, but the core authentication mechanisms (MFA, strong credentials) do not apply because the session is already authenticated when hijacking occurs; the technique can still be executed against any active session within the allowed window.
- T1566partialprevents — A.8.5's MFA requirement prevents credential-phishing from yielding unauthorized access even when a user surrenders a password, but phishing that delivers malicious attachments, links to drive-by exploits, or social-engineering victims into installing remote-access tools is not addressed by authentication controls, so only a slice of the technique is prevented.
- T1589.001partialdetects — A.8.5 explicitly requires logging of successful/unsuccessful log-on attempts, raising security events/alerts on potential breaches or excessive failures, which surfaces credential-gathering activity (e.g. brute-force, phishing attempts, or anomalous logons) when it interacts with protected systems; this is only a slice of the broad pre-compromise gathering methods in T1589.001 such as dark-web purchases, breach dumps, or site compromises that never touch the victim's authentication surface.
- T1599partialprevents — A.8.5's strong authentication, MFA, and brute-force protections raise the bar against credential-based compromise of boundary devices, but T1599 also succeeds via vulnerability exploitation and configuration abuse of network devices, which A.8.5 does not address; the technique's mechanism is broader than the authentication path A.8.5 closes.
- T1602.001partialprevents — A.8.5's requirements for strong authentication and brute-force protection (item e) would, if applied to SNMP, push deployments toward SNMPv3 with credentials and away from guessable community strings, blocking some unauthorized MIB queries; however, the clause is framed around user logon to systems/applications and does not specifically address network device management protocols, so implementations fully conformant with A.8.5 can still expose SNMP with weak or default community strings, leaving the technique reachable.
- T1602.002partialprevents — A.8.5's authentication-strength requirements (MFA, brute-force protection, credential hygiene) raise the bar for accessing network device management interfaces used in T1602.002 (SNMP, Smart Install), but the technique also exploits weak/default community strings, hardcoded credentials, and protocol-level weaknesses that authentication strength alone does not eliminate, and many network devices cannot support MFA.
- T1621partialdetects — A.8.5's logging of authentication attempts (f) and raising of security events on suspicious log-on patterns (g) would surface the repeated authentication requests characteristic of MFA fatigue/push bombing, but the clause's detection scope is the log-on event itself, not the broader pattern of MFA prompt abuse or the user's acceptance of the prompt, which is the technique's core mechanism.
- T1621partialprevents — A.8.5 mandates MFA for critical systems and requires additional factors under specific circumstances, which directly addresses the MFA fatigue vector, but the control sets a requirement rather than a universal mechanism and cannot prevent an authorized user from accepting a push notification they believe is legitimate.
- T1669partialprevents — A.8.5 requires strong authentication including MFA for critical systems, which can prevent unauthorized access to secured Wi-Fi networks, but the control does not address open Wi-Fi networks or physical proximity attacks that bypass authentication entirely.
- T1684.002partialprevents — A.8.5 mandates strong authentication and MFA for critical systems, which can prevent account compromise that enables spoofing, but it does not enforce the email authentication protocols (SPF, DKIM, DMARC) that directly prevent header spoofing.
Prevented OWASP Web Top 10 (2025) risks (11)
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
- A07mostlyblocks — A.8.5 directly mandates strong/multi-factor authentication techniques, log-on hardening (no info leakage, brute-force protection, no cleartext transmission, session termination), and MFA for critical systems, which squarely blocks the dominant authentication-bypass, credential-stuffing, weak-reset, and session-hijacking patterns named in A07:2025; residual gaps (e.g., implementation flaws in the chosen technique or non-covered edge cases) keep it from full.
- A10partialblocks — A.8.5's log-on procedure rules (a–e, g–j) directly block information leaks, brute-force, cleartext exposure and fail-open authentication paths that are core to the A10 class; residual logic-flaw error handling in application code and non-authentication exception paths remain 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.