Recognition-decision-execution: a new design principle for intelligent antimicrobial systems

The continued rise of antimicrobial resistance highlights fundamental limitations of prevailing antibacterial strategies, which remain largely dominated by continuous, non-directed killing. Although combination therapies and multifunctional materials are frequently described as synergistic, many reported effects primarily reflect pharmacokinetic enhancement or local concentration increases, rather than genuine reprogramming of antibacterial behavior. As a result, most current approaches still impose sustained selective pressure and offer limited control over timing, location, and intensity of killing. Here, we propose the Recognition-Decision-Execution (RDE) framework as a function-driven organizational principle for antibacterial system design. In this framework, antibacterial activity is decomposed into three orthogonal yet logically coupled modules: pathogen recognition, decision-making, and killing execution. Recognition modules provide contextual information regarding pathogen identity, microenvironmental state, and infection dynamics; decision modules interpret these inputs to authorize action with defined spatiotemporal and intensity constraints; and execution modules eliminate pathogens only upon conditional approval. This architecture transforms antibacterial activity from a default, always-on process into a programmable, event-driven intervention that may reshape rather than universally reduce selective pressure. We discuss how functional decoupling enables a clear distinction between apparent (pseudo-) synergy and true mechanistic synergy, and how regulatory-execution interactions can qualitatively reshape killing mechanisms rather than merely amplifying them. Importantly, functional roles are not intrinsic to specific agents: antimicrobial peptides, nanomaterials, and nucleic acid aptamers can dynamically interchange between regulatory and execution functions depending on system architecture. The aptamer-nanomaterial-antimicrobial peptide triad serves as a minimal closed-loop example illustrating how sensing, decision-making, and execution can be independently optimized yet synergistically integrated. By emphasizing system logic, modular verifiability, and conditional authorization of killing, the RDE framework offers a unifying conceptual language for the design of programmable and adaptive antibacterial therapies. It should be emphasized that the RDE framework represents a unifying conceptual paradigm and set of testable design hypotheses. Most reported antibacterial examples only achieve partial pairwise functional slices rather than fully-validated ternary closed-loop RDE architectures.