We present WebParserHunter, an interactive web application for automatically identifying parsing and validation functions in x86-32 and x86-64 binary executables without source code. This demo is the companion of the ParserHunter methodology [4] that relies on the SAFE embedding model and Graph Sage GNNs for code classification: analysts can upload any binary executable to the application and receive a ranked list with predicted probabilities of it being a parser, along with CFG visualizations and inline assembly. They can then launch an angr-based symbolic execution or a fuzzing campaign for deeper auditing within the same browser session.

Identifying Parser Functions in Binary Executables with WebParserHunter / Scapin, M., Pinelli, F., Galletta, L.. - 16950:(2026), pp. 320-323. (European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2026 Naples, Italy 7-11 September, 2026) [10.1007/978-3-032-37685-5_26].

Identifying Parser Functions in Binary Executables with WebParserHunter

Scapin, Marco
;
Pinelli, Fabio;Galletta, Letterio
2026

Abstract

We present WebParserHunter, an interactive web application for automatically identifying parsing and validation functions in x86-32 and x86-64 binary executables without source code. This demo is the companion of the ParserHunter methodology [4] that relies on the SAFE embedding model and Graph Sage GNNs for code classification: analysts can upload any binary executable to the application and receive a ranked list with predicted probabilities of it being a parser, along with CFG visualizations and inline assembly. They can then launch an angr-based symbolic execution or a fuzzing campaign for deeper auditing within the same browser session.
2026
9783032376848
9783032376855
Binary analysis, Parser identification, Graph Neural, Networks, Security auditing, Symbolic execution
File in questo prodotto:
File Dimensione Formato  
ecml_pkdd_demo_phunter.pdf

Accesso aperto

Tipologia: Documento in Pre-print
Licenza: Creative commons
Dimensione 454.5 kB
Formato Adobe PDF
454.5 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11771/44318
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • OpenAlex 0
social impact