Know the threat.
Define the boundary.
Nineteen threat and assurance topics. Each starts with the protected decision and makes complementary controls explicit.
Explore the attack surface.
AiTM phishing & MFA bypass↗
Bind authority to the legitimate device, beyond the login ceremony.
Mechanism & boundaryAccount takeover↗
Examine what an attacker can do after acquiring a credential.
Mechanism & boundaryToken theft & replay↗
A copied token should not become portable authority.
Mechanism & boundaryAgent tool-call abuse↗
Constrain which tools an agent may invoke, on which resources.
Mechanism & boundaryBroken access control↗
Verify the object, tenant and action at the execution boundary.
Mechanism & boundaryThird-party script injection↗
Understand the boundary around the browser signing context.
Mechanism & boundaryAI-assisted intrusion↗
Protect sensitive actions from misuse of session authority.
Mechanism & boundaryRansomware & AI exploitation↗
Contain session authority within a broader incident response.
Mechanism & boundaryPrompt injection↗
Separate untrusted content from authority to act.
Mechanism & boundaryAI session hijacking↗
Keep session authority bound to its intended context.
Mechanism & boundarySession injection & replay↗
Reject reused or incorrectly bound protected requests.
Mechanism & boundaryHarvest now, decrypt later↗
Plan confidentiality for information with a long useful life.
Assurance topicOWASP control mapping↗
Map mechanisms to specific application risks.
Mechanism & boundarySession fixation↗
Re-establish authority across authentication transitions.
Mechanism & boundaryIndirect prompt injection & RAG poisoning↗
Constrain actions influenced by retrieved content.
Mechanism & boundaryNon-human identity abuse↗
Make workload authority explicit, bounded and revocable.
Mechanism & boundaryBGP & DNS hijacking↗
Separate endpoint authenticity from routing assurance.
Mechanism & boundaryMetadata & lawful intercept↗
Inspect what remains visible beyond encrypted content.
Assurance topicPost-quantum migration↗
Treat authenticity and confidentiality as distinct migration tasks.
Assurance topicWhere each threat topic meets the frameworks.
Cross-references for the threat topics above. Open a framework to see which topics relate to it, with a link to the official source.
References show where a topic relates to a framework, standard or regulation. They are not a claim of compliance, certification or control coverage, and not legal advice. The mapping is a working draft that has not been independently reviewed. Check the current text of each source.
Standards and frameworks
US standards and guidance for security, digital identity, zero trust and cryptography.
Threat topics it relates to
- Account takeoverSP 800-63 Digital Identity Guidelines · SP 800-207 Zero Trust Architecture
- AI-assisted intrusionAI Risk Management Framework · SP 800-207 Zero Trust Architecture
- AI session hijackingSP 800-207 Zero Trust Architecture · SP 800-63 Digital Identity Guidelines · AI Risk Management Framework
- AiTM phishing & MFA bypassSP 800-63 Digital Identity Guidelines
- BGP & DNS hijackingSP 800-189 BGP security and DDoS mitigation
- Broken access controlSP 800-207 Zero Trust Architecture
- Agent tool-call abuseAI Risk Management Framework · SP 800-207 Zero Trust Architecture
- Harvest now, decrypt laterFIPS 203 (ML-KEM) · FIPS 204 (ML-DSA) · FIPS 205 (SLH-DSA) · Post-quantum cryptography project
- Indirect prompt injection & RAG poisoningAI Risk Management Framework
- Non-human identity abuseSP 800-207 Zero Trust Architecture
- Token theft & replaySP 800-63 Digital Identity Guidelines · SP 800-207 Zero Trust Architecture
- OWASP control mappingSP 800-218 Secure Software Development Framework
- Post-quantum migrationFIPS 203 (ML-KEM) · FIPS 204 (ML-DSA) · FIPS 205 (SLH-DSA) · Post-quantum cryptography project
- Prompt injectionAI Risk Management Framework
- Ransomware & AI exploitationSP 800-207 Zero Trust Architecture
- Session fixationSP 800-63 Digital Identity Guidelines
- Session injection & replaySP 800-63 Digital Identity Guidelines · SP 800-207 Zero Trust Architecture
- Third-party script injectionSP 800-218 Secure Software Development Framework
International information security management standard. Annex A lists the controls.
Threat topics it relates to
- Account takeoverAnnex A.8.5 Secure authentication
- AI-assisted intrusionAnnex A.8.16 Monitoring activities
- AI session hijackingAnnex A.8.5 Secure authentication
- AiTM phishing & MFA bypassAnnex A.8.5 Secure authentication
- BGP & DNS hijackingAnnex A.8.20 Networks security
- Broken access controlAnnex A.8.3 Information access restriction
- Agent tool-call abuseAnnex A.5.15 Access control
- Harvest now, decrypt laterAnnex A.8.24 Use of cryptography
- Metadata & lawful interceptAnnex A.8.24 Use of cryptography
- Non-human identity abuseAnnex A.5.16 Identity management
- Token theft & replayAnnex A.8.5 Secure authentication
- Post-quantum migrationAnnex A.8.24 Use of cryptography
- Ransomware & AI exploitationAnnex A.8.16 Monitoring activities
- Session fixationAnnex A.8.5 Secure authentication
- Session injection & replayAnnex A.8.5 Secure authentication
- Third-party script injectionAnnex A.5.21 ICT supply chain security
AICPA assurance reporting against the Trust Services Criteria.
Threat topics it relates to
- Account takeoverCC6 Logical and physical access controls
- AI-assisted intrusionCC6 Logical and physical access controls
- Broken access controlCC6 Logical and physical access controls
- Agent tool-call abuseCC6 Logical and physical access controls
- Non-human identity abuseCC6 Logical and physical access controls
- Token theft & replayCC6 Logical and physical access controls
- Ransomware & AI exploitationCC7 System operations
- Third-party script injectionCC9 Risk mitigation
Security standard for organisations that handle payment card data.
Threat topics it relates to
- Account takeoverv4.0 Requirement 8 Identify users and authenticate access
- AiTM phishing & MFA bypassv4.0 Requirement 8 Identify users and authenticate access
- Broken access controlv4.0 Requirement 7 Restrict access by need to know
- Non-human identity abusev4.0 Requirement 8 Identify users and authenticate access
- Token theft & replayv4.0 Requirement 8 Identify users and authenticate access
Prioritised safeguards from the Center for Internet Security (version 8).
Threat topics it relates to
- Account takeoverControl 6 Access control management
- AiTM phishing & MFA bypassControl 6 Access control management
- Broken access controlControl 6 Access control management
- Non-human identity abuseControl 5 Account management
- Ransomware & AI exploitationControl 11 Data recovery
Community risk lists and guidance for web, API and LLM application security.
Threat topics it relates to
- Account takeoverTop 10 A07 Identification and Authentication Failures · Session Management Cheat Sheet
- AI-assisted intrusionTop 10 for LLM Applications
- AI session hijackingSession Management Cheat Sheet
- AiTM phishing & MFA bypassTop 10 A07 Identification and Authentication Failures
- Broken access controlTop 10 A01 Broken Access Control · API Security Top 10 API1 Broken Object Level Authorization
- Agent tool-call abuseTop 10 for LLM Applications
- Indirect prompt injection & RAG poisoningTop 10 for LLM Applications
- Non-human identity abuseTop 10 A07 Identification and Authentication Failures
- Token theft & replayTop 10 A07 Identification and Authentication Failures · Session Management Cheat Sheet
- OWASP control mappingTop 10 A01 Broken Access Control · Top 10 A07 Identification and Authentication Failures · Top 10 A08 Software and Data Integrity Failures · API Security Top 10 API1 Broken Object Level Authorization · Top 10 for LLM Applications · Session Management Cheat Sheet
- Prompt injectionTop 10 for LLM Applications
- Session fixationSession Management Cheat Sheet · Top 10 A07 Identification and Authentication Failures
- Session injection & replaySession Management Cheat Sheet · Top 10 A07 Identification and Authentication Failures
- Third-party script injectionTop 10 A08 Software and Data Integrity Failures
US NSA guidance on quantum-resistant algorithms for national security systems.
Threat topics it relates to
- Harvest now, decrypt laterNSA post-quantum cybersecurity resources (CNSA 2.0)
- Post-quantum migrationNSA post-quantum cybersecurity resources (CNSA 2.0)
Regulation
EU data-protection regulation, including security of processing.
Threat topics it relates to
- Account takeoverArt. 32 Security of processing
- AI-assisted intrusionArt. 32 Security of processing
- AI session hijackingArt. 32 Security of processing
- BGP & DNS hijackingArt. 32 Security of processing
- Broken access controlArt. 32 Security of processing
- Harvest now, decrypt laterArt. 32 Security of processing
- Metadata & lawful interceptArt. 5 Principles for processing · Art. 32 Security of processing
- Ransomware & AI exploitationArt. 32 Security of processing
EU regulation on digital operational resilience for financial entities.
Threat topics it relates to
- Account takeoverArt. 9 Protection and prevention
- AiTM phishing & MFA bypassArt. 9 Protection and prevention
- BGP & DNS hijackingArt. 9 Protection and prevention
- Harvest now, decrypt laterArt. 9 Protection and prevention
- Post-quantum migrationArt. 9 Protection and prevention
- Ransomware & AI exploitationArt. 9 Protection and prevention
- Third-party script injectionChapter V ICT third-party risk
EU regulation on electronic identification and trust services.
Threat topics it relates to
- Account takeoverRegulation (EU) No 910/2014
- AiTM phishing & MFA bypassRegulation (EU) No 910/2014
Threat knowledge bases
MITRE ATT&CK and ATLAS
Official source (opens the official source in a new tab)Catalogues of adversary techniques against enterprise systems (ATT&CK) and AI systems (ATLAS).
Threat topics it relates to
- Account takeoverATT&CK T1078 Valid Accounts · ATT&CK T1539 Steal Web Session Cookie
- AI-assisted intrusionATT&CK T1078 Valid Accounts
- AI session hijackingATT&CK T1539 Steal Web Session Cookie · ATT&CK T1550 Use Alternate Authentication Material
- AiTM phishing & MFA bypassATT&CK T1557 Adversary-in-the-Middle · ATT&CK T1566 Phishing
- Indirect prompt injection & RAG poisoningATLAS AML.T0051 LLM Prompt Injection
- Non-human identity abuseATT&CK T1078 Valid Accounts · ATT&CK T1528 Steal Application Access Token
- Token theft & replayATT&CK T1528 Steal Application Access Token · ATT&CK T1550 Use Alternate Authentication Material
- Prompt injectionATLAS AML.T0051 LLM Prompt Injection
- Ransomware & AI exploitationATT&CK T1486 Data Encrypted for Impact
- Session injection & replayATT&CK T1550 Use Alternate Authentication Material
- Third-party script injectionATT&CK T1195 Supply Chain Compromise
See the session lifecycle under test.
Explore the boundary. Inspect the decision. Trace the recovery.