Invariant-Centered, Agent-Assisted Defense: A Security Architecture for Unbounded Attack Techniques and Unenumerable Attack Objectives

Ajayi Abisoye; Nurudeen Hussain; Abolaji Adebayo

doi:10.56201/ijemt.vol.12.no7.2026.pg193.207

📄 Download Paper

Invariant-Centered, Agent-Assisted Defense: A Security Architecture for Unbounded Attack Techniques and Unenumerable Attack Objectives

Ajayi Abisoye, Nurudeen Hussain, Abolaji Adebayo

Abstract

Generative AI has industrialized the attack on human judgment: voice cloning, real-time video synthesis, and hyper-personalized phishing convert social engineering from a craft into a commodity, while a residual category of attacks — those that redefine the objective itself, as ransomware once did — cannot be enumerated in advance. We propose a defensive architecture for these two conditions built on a single organizing claim: under unbounded attack techniques and unenumerable attack objectives, the highest-value security investments are those that hold regardless of technique. The architecture layers a mesh of strictly advisory AI sentinel agents, which detect cross-surface incoherence rather than synthetic content, above a deliberately simple, non-interpretive enforcement core holding a small set of hard invariants that no intelligence — human or artificial — can rewrite quickly. We give the reactive-defender objection a full treatment: the claim that any AI-era defense necessarily responds late. Our answer is that the objection is correct for every layer that must recognize attacks, and that the architecture is designed around that concession — detection layers are built for their own capture, while the invariant core does not race because it constrains consequences rather than recognizing techniques. The framework relocates residual reactivity to a single measurable point, the interval between the world changing and the invariant set catching up, and proposes that interval as the headline resilience metric. We state residual risks explicitly and outline an evaluation agenda.

Keywords

social engineeringgenerative AIdeepfakessecurity architectureprompt injectionzero trustresilience engineeringhuman factorsinvariants

References

[1] Saltzer, J.H., and Schroeder, M.D. (1975). The Protection of Information in Computer Systems. Proceedings of the IEEE, 63(9). (Least privilege, economy of mechanism, fail-safe defaults, separation of privilege.) [2] Thompson, K. (1984). Reflections on Trusting Trust. Communications of the ACM, 27(8). [3] Hardy, N. (1988). The Confused Deputy (or why capabilities might have been invented). ACM SIGOPS Operating Systems Review, 22(4). [4] Rose, S., Borchert, O., Mitchell, S., and Connelly, S. (2020). Zero Trust Architecture. NIST Special Publication 800-207. [5] Parasuraman, R., and Riley, V. (1997). Humans and Automation: Use, Misuse, Disuse, Abuse. Human Factors, 39(2). [6] Reason, J. (1990). Human Error. Cambridge University Press. (Defense-in-depth and the Swiss cheese model of accident causation.) [7] Klein, G., et al. (2009). seL4: Formal Verification of an OS Kernel. Proceedings of the 22nd ACM Symposium on Operating Systems Principles . [8] MITRE ATT&CK. Enterprise Matrix: tactics and techniques knowledge base. https://attack.mitre.org [9] Vassilev, A., et al. (2024). Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations. NIST AI 100-2 series. [10] OWASP Foundation. OWASP Top 10 for Large Language Model Applications. (LLM01: Prompt Injection.) https://owasp.org [11] Cialdini, R. (1984). Influence: The Psychology of Persuasion. (Authority, urgency/scarcity, social proof as stable levers of compliance.) [12] Verizon. Data Breach Investigations Report , recent annual editions. (Human element implicated in a majority of analyzed breaches.) [13] Federal Bureau of Investigation, Internet Crime Complaint Center (IC3). Annual Internet Crime Reports. (Business email compromise loss figures.) [14] Widely reported incident: Arup (Hong Kong office), early 2024 — employee authorized transfers of approximately USD 25 million following a video conference in which participants were AI-generated deepfakes of colleagues. To be replaced with a specific press citation before publication. [15] Cybersecurity and Infrastructure Security Agency (2022). Implementing Phishing- Resistant MFA. (FIDO2/WebAuthn guidance.) [16] Hollnagel, E., Woods, D.D., and Leveson, N. (eds.) (2006). Resilience Engineering: Concepts and Precepts. Ashgate. See also Leveson, N. (2011). Engineering a Safer World. MIT Press. [17] Taleb, N.N. (2012). Antifragile: Things That Gain from Disorder. Random House. [18] Geer, D., et al. (2003). CyberInsecurity: The Cost of Monopoly. Computer & Communications Industry Association report. (Software monoculture risk.) [19] Avizienis, A. (1985). The N-Version Approach to Fault-Tolerant Software. IEEE Transactions on Software Engineering, SE-11(12). Note on citations: references [1]–[13] and [15]–[19] are canonical, verifiable sources; edition- specific figures (e.g., DBIR percentages, IC3 totals) should be confirmed against the latest annual releases before external publication. This draft was prepared with AI assistance and reviewed for source accuracy at the level of claims made; it has not undergone independent fact-checking.

Invariant-Centered, Agent-Assisted Defense: A Security Architecture for Unbounded Attack Techniques and Unenumerable Attack Objectives

Abstract

Keywords

References

More Articles from INTERNATIONAL JOURNAL OF ENGINEERING AND MODERN TECHNOLOGY