mlsanityData to 7 Oct 2026

Identifying Adversary Tactics and Techniques in Malware Binaries with an LLM Agent

arXiv 2602.06325 ↗First posted 6 Feb 2026cs.CR cs.SEcited by 2

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

Understanding TTPs (Tactics, Techniques, and Procedures) in malware binaries is essential for security analysis and threat intelligence, yet remains challenging in practice. Real-world malware binaries are typically stripped of symbols, contain large numbers of functions, and distribute malicious behavior across multiple code regions, making TTP attribution difficult. Recent large language models (LLMs) offer strong code understanding capabilities, but applying them directly to this task faces challenges in identifying analysis entry points, reasoning under partial observability, and misalignment with TTP-specific decision logic. We present TTPDetect, the first LLM agent for recognizing TTPs in stripped malware binaries. TTPDetect combines dense retrieval with LLM-based neural retrieval to narrow the space of analysis entry points. TTPDetect further employs a function-level analyzing agent consisting of a Context Explorer that performs on-demand, incremental context retrieval and a TTP-Specific Reasoning Guideline that achieves inference-time alignment. We build a new dataset that labels decompiled functions with TTPs across diverse malware families and platforms. TTPDetect achieves 93.25% precision and 93.81% recall on function-level TTP recognition, outperforming baselines by 10.38% and 18.78%, respectively. When evaluated on real world malware samples, TTPDetect recognizes TTPs with a precision of 87.37%. For malware with expert-written reports, TTPDetect recovers 85.7% of the documented TTPs and further discovers, on average, 10.5 previously unreported TTPs per malware.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 19 of its 53 references that are in the map. First posted on arXiv 6 Feb 2026.

Builds on

19 of 53 references.

References · 19

  1. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,383
  2. 2018TESSERACT: Eliminating Experimental Bias in Malware Classification across Space and Timecited by 108
  3. 2022Partially Observable Markov Decision Processes in Robotics: A Surveycited by 81
  4. 2025From Hard Refusals to Safe-Completions: Toward Output-Centric Safety Trainingcited by 49
  5. 2025RepoAudit: An Autonomous LLM-Agent for Repository-Level Code Auditingcited by 47
  6. 2019An Observational Investigation of Reverse Engineers' Processescited by 20
  7. 2025LAMD: Context-driven Android Malware Detection and Classification with LLMscited by 18
  8. 2025On Benchmarking Code LLMs for Android Malware Analysiscited by 12
  9. 2018Context2Name: A Deep Learning-Based Approach to Infer Natural Variable Names from Usage Contextscited by 11
  10. 2024TREC: APT Tactic / Technique Recognition via Few-Shot Provenance Subgraph Learningcited by 6
  11. 2024Tactics, Techniques, and Procedures (TTPs) in Interpreted Malware: A Zero-Shot Generation with Large Language Modelscited by 6
  12. 2016Automated Synthesis of Semantic Malware Signatures using Maximum Satisfiabilitycited by 5
7 more
  1. 2023Symbol Preference Aware Generative Models for Recovering Variable Names from Stripped Binarycited by 5
  2. 2025Large Language Models for Validating Network Protocol Parserscited by 5
  3. 2025Is "Knowing It's Malicious Enough?" Evaluating LLMs for Fine-Grained Malware Behavior Auditingcited by 3
  4. 2025On Technique Identification and Threat-Actor Attribution using LLMs and Embedding Modelscited by 2
  5. 2025Sense and Sensitivity: Examining the Influence of Semantic Recall on Long Context Code Understandingcited by 2
  6. 2025MalLoc: Toward Fine-grained Android Malicious Payload Localization via LLMscited by 2
  7. 2025DroidTTP: Mapping Android Applications with TTP for Cyber Threat Intelligencecited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Malaika: Understanding Malware through Tri-Grounded Agentic Reasoningcited by 0
  2. 2026MARS: Malware Analysis with Rule-Based Scoring of LLM Claimscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Advancing TTP Analysis: Harnessing the Power of Large Language Models with Retrieval Augmented Generation
  2. 2026AsmRAG: LLM-Driven Malware Detection by Retrieving Functionally Similar Assembly Code
  3. 2026ContextLeak: Exfiltrating LLM Agent Context via Malicious Tools
  4. 2025XGen-Q: An Explainable Domain-Adaptive LLM Framework with Retrieval-Augmented Generation for Software Security
  5. 2026ALIBI: Adversarial Legitimacy Injection in Binary Input against LLM Malware Analyzers
  6. 2026CHASE: LLM Agents for Dissecting Malicious PyPI Packages
  7. 2026MalTool: Malicious Tool Attacks on LLM Agents
  8. 2026When the Manual Lies: A Realistic Benchmark to Evaluate MCP Poisoning Attacks for LLM Agents