Newest papers

Typed Template Fuzzing

Lu Maltsis · 2026 · ACM SIGPLAN International Conference on Systems, Programming, Languages, and Applications: Software for Humanity (SPLASH)

DOI: 10.1145/3837729.3839574 · cited by 0 works (OpenAlex) · open-access PDF

Korean

Fuzzing는 높은 양의 입력으로 소프트웨어 행동을 테스트함으로써 버그를 발견합니다. 다른 타겟을 위해 정의적으로 유효한 입력을 만들려면 전통적인 fuzzers는 상당한 엔지니어링 노력을 필요로합니다. LLM 기반 fuzzing 접근법은 다양한 소프트웨어에 대한 입력을 만들 수 있지만 새로운 문제를 도입합니다. 각 테스트는 전통적인 fuzzing의 속도를 희생하여 LLM 쿼리를 필요로합니다. 우리는 Typed Template Fuzzing (TTF)를 도입합니다. 그것은 사전 처리 단계에서 LLM을 사용하여 씨앗 파일의 정의 정보를 인코딩하는 템플릿을 만들 수 있습니다. 이러한 템플릿은 정의적으로 인식되는 돌연변이를 가능하게합니다. 각 테스트는 분석기, 유형 검사기

Machine translation (Meta M2M100); the English original follows.

Abstract

Fuzzing finds bugs by testing software behavior with high volumes of input. To create semantically valid inputs for different targets, traditional fuzzers require significant engineering effort. LLM-based fuzzing approaches can be adapted to create inputs for a wide range of software but introduce new issues: every test needs an LLM query, sacrificing the speed of traditional fuzzing. We introduce Typed Template Fuzzing (TTF). It uses LLMs in a preprocessing step to create templates which encode semantic information of seed files. These templates enable semantically-aware mutations without the need for a parser, type-checker, or further LLM queries. We have implemented TTF in our tool Typos. Using Typos we have run a five-month fuzzing campaign, finding 97 previously undiscovered bugs. Typos achieves a 94.7% rate of valid inputs across a wide range of formats including SMT-LIB2, C, C++, Go, Rust, SystemVerilog, Scala, Java, and domain-specific libraries such as PyTorch and Qiskit.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Fuzz testing, Computer science, Preprocessor, Template, Software, Programming language

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31342, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange