oss-fuzz-gen and LLMFuzzer
These are complements: OSS-Fuzz-Gen generates fuzzing inputs for traditional software using LLMs, while LLMFuzzer uses fuzzing techniques to test the LLMs themselves, addressing different layers of the testing pipeline.
About oss-fuzz-gen
google/oss-fuzz-gen
LLM powered fuzzing via OSS-Fuzz.
Generates fuzz targets across C/C++, Java, and Python by prompting multiple LLMs (Vertex AI, Gemini, GPT-4) to synthesize harnesses, then evaluates them against OSS-Fuzz infrastructure using metrics like compilability, crash detection, and coverage gain versus human-written targets. The framework discovered 30+ real vulnerabilities including CVE-2024-9143 in OpenSSL, demonstrating LLM-generated targets can achieve up to 29% additional line coverage on production projects.
About LLMFuzzer
mnns/LLMFuzzer
🧠 LLMFuzzer - Fuzzing Framework for Large Language Models 🧠 LLMFuzzer is the first open-source fuzzing framework specifically designed for Large Language Models (LLMs), especially for their integrations in applications via LLM APIs. 🚀💥
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work