Skip to content

Back to the catalog

tmylla/Awesome-LLM4Cybersecurity
github.comrepo

tmylla/Awesome-LLM4Cybersecurity

Awesome-LLM4Cybersecurity is a GitHub-hosted, continuously updated literature review (861+ papers as of 2026) organized around three research questions: building domain-specific security LLMs, applying LLMs to cybersecurity tasks (threat intelligence, fuzzing, vulnerability detection, program repair, insecure code generation, LLM-assisted attack and defense), and future directions like security-focused LLM agents. It accompanies the paper "When LLMs Meet Cybersecurity: A Systematic Literature Review" (Zhang et al., Cybersecurity journal, 2025) and serves as an index for researchers tracking the intersection of large language models and security.

TL;DR Continuously updated GitHub literature index of 861+ papers on applying LLMs to cybersecurity tasks like vulnerability detection, fuzzing, threat intel, and LLM-based attack/defense agents.
Open original ↗
Document

What it is

Awesome-LLM4Cybersecurity is a GitHub reading list of 861+ papers on large language models and security. It accompanies Zhang et al., "When LLMs Meet Cybersecurity: A Systematic Literature Review" (Cybersecurity, vol. 8, no. 1, 2025). The README is an index. The full paper list sits in a separate linked file.

Coverage by category

The per-category counts add up to 861 and show where the literature is concentrated:

Research questionCategoryPapers
RQ1: Domain LLMsCybersecurity evaluation benchmarks93
RQ1: Domain LLMsFine-tuned domain LLMs37
RQ2: ApplicationsLLM-assisted defense150
RQ2: ApplicationsLLM-assisted attack126
RQ2: ApplicationsVulnerability detection120
RQ2: ApplicationsProgram/vulnerability repair78
RQ2: ApplicationsOthers72
RQ2: ApplicationsThreat intelligence54
RQ2: ApplicationsInsecure code generation34
RQ2: ApplicationsFuzzing26
RQ3: Future directionsAgent4Cybersecurity71

Fuzzing and insecure code generation are the thinnest lists. Defense, attack and vulnerability detection are the largest.

Update history

Recent updates have been large. The README logs these additions:

  • August 2026: 105 papers, coverage through 20 August 2026.
  • June 2026: 108 papers.
  • February 2026: 80 papers, covering August 2025 to January 2026.
  • November 2025: 176 papers, covering March to August 2025.

Earlier updates ran from September 2024 (75 papers) through March 2025 (33 papers). Since January 2025 each paper entry includes its publication venue.

Caveats

The README gives no inclusion criteria, quality ratings or per-paper summaries. Whether a given paper is listed has to be checked in the linked full file. The most recent coverage date is 20 August 2026, so later work is missing.

Technologies named here

Related in AI Security