Advanced System-Level Fuzzing for OpenPrinting: Deep State Exploration and LLM-Augmented Mutation

This project aims to transition OpenPrinting’s security infrastructure from fragmented unit-testing to a comprehensive, state-aware system fuzzing framework. While current fuzzers target isolated helper functions, they fail to exercise the deep implementation logic of complex media parsers (PDF, PCL, Raster). Utilizing AFL++, Honggfuzz, and OSS-Fuzz-Gen, I will develop API-Sequence-Based harnesses that mimic authentic user-triggered scenarios. Key deliverables include a unified consumption shim for the media pipeline, IR-based recombination algorithms for nested object abstractions, and a curated high-quality seed corpus. My preliminary discovery of a zero-day vulnerability in ipp.c (GHSA-67hg-386m-x83h) underscores the critical need for this systemic approach to harden the open-source printing ecosystem.