LLM-Based Agents for Software and Systems Security: Approaches, Applications, and Assessment
Organizations: University at Buffalo, SUNY, Buffalo, New York, USA
Abstract
Software and systems security workflows are typically procedural: analysts inspect heterogeneous artifacts, form hypotheses, invoke tools, interpret outputs, and revise plans. Large language model (LLM)-based agents, which can plan, use tools, retain state, and revise actions across multi-step workflows, are being rapidly adopted to automate this work. Given the consequences of delegating security decisions to autonomous systems, understanding how such agents are built, used, and assessed is crucial. Yet to this date, there remains a lack of systematic understanding of what has been done and how far we are in this field: the term "agent" is applied inconsistently, applications differ sharply in risk, and assessment protocols are often incomparable. To gain a comprehensive and coherent view of this area hence inform relevant future research, this paper provides a systematic literature review of the (1) technical approaches, including agent architecture, perception, memory, reasoning and planning, action space, orchestration, and self-improvement, (2) applications, with respect to the security tasks served, and (3) assessment, including the datasets, outcome and trajectory metrics, safety measures, and baselines considered, over the peer-reviewed literature spanning the emergence of this area (2023--2026). Our synthesis reveals a field that has built agents able to act but not yet agents whose authority is bounded or whose behavior is auditable. In addition to knowledge systematization, we also extend our insights into the limitations of and challenges faced by current approach, application, and assessment designs, which shed light on potentially promising future research directions.
Figures & tables
| Survey | Scope | Agents | Instrument | Approach | Application | Assessment | Year |
| Xu et al. ( Xu et al., 2025a ) | LLMs for cybersecurity | ○ | ◐ | ◐ | ● | ◐ | 2024 |
| Zhou et al. ( Zhou et al., 2025 ) | LLMs for vuln. detect./repair | ○ | ◐ | ◐ | ◐ | ◐ | 2024 |
| Fan et al. ( Fan et al., 2023 ) | LLMs for SE | ○ | ○ | ◐ | ○ | ○ | 2023 |
| Liu et al. ( Liu et al., 2026 ) | LLM agents for SE | ● | ○ | ● | ○ | ◐ | 2024 |
| Xu et al. ( Xu et al., 2025b ) | Agents for cyberattacks | ● | ● | ◐ | ◐ | ○ | 2025 |
| Hassanin et al. ( Hassanin and Moustafa, 2024 ) | LLMs for cyber defence | ○ | ◐ | ○ | ◐ | ○ | 2024 |
| Aspect | Attribute | Item | Description |
|---|---|---|---|
| Approach | Agent architecture | Agent multiplicity | Single vs. multi-agent organization. |
| Internal structure | Module, pipeline, guardrail, fine-tuning, or hybrid layout. | ||
| Backbone LLM role | LLM as actor, planner, verifier, critic, or controller. | ||
| Agent perception | Input sources | Code, binaries, traces, logs, environments, images, or standards. | |
| Perception mechanisms | Ingest, retrieve, summarize, embed, or observe via tools. | ||
| Agent memory | Memory types | In-context, external, long-term, graph, or parametric memory. |
| Type | Acronym | Description | # Papers |
| Conference | TrustCom | IEEE International Conference on Trust, Security and Privacy in Computing and Communications | 5 |
| USENIX Security | USENIX Security Symposium | 4 | |
| ICSE | IEEE/ACM International Conference on Software Engineering | 3 | |
| ICAIC | IEEE International Conference on AI in Cybersecurity | 3 | |
| BigData | IEEE International Conference on Big Data | 3 | |
| MILCOM | IEEE Military Communications Conference | 3 |
LLM Agent Security
Large language model agent security studies attacks and safeguards for systems that plan, use tools, retain memory, and act across applications.
See all 327 papers