Inna Skarbovsky

31 papers C 1Journal 12Unranked 17
YearRankTypeTitle / Venue / Authors
2025 J jnl
CoRR
Dany Moshkovich, Hadar Mulian, Sergey Zeltyn, Natti Eder, Inna Skarbovsky, Roy Abitbol
2025 J jnl
Data Knowl. Eng.
Dirk Fahland, Fabiana Fournier, Lior Limonad, Inna Skarbovsky, Ava J. E. Swevels
2025 J jnl
Künstliche Intell.
Fabiana Fournier, Lior Limonad, Inna Skarbovsky, Yuval David
2025 conf
BPM (Forum)
Yuval David, Fabiana Fournier, Lior Limonad, Inna Skarbovsky
2025 J jnl
CoRR
Yuval David, Fabiana Fournier, Lior Limonad, Inna Skarbovsky
2024 J jnl
CoRR
Dirk Fahland, Fabiana Fournier, Lior Limonad, Inna Skarbovsky, Ava J. E. Swevels
2024 J jnl
CoRR
Lior Limonad, Fabiana Fournier, Juan Manuel Vera, Inna Skarbovsky, Shlomit Gur, Raquel Lazcano
2024 conf
Business Process Management Workshops
Fabiana Fournier, Lior Limonad, Inna Skarbovsky
2024 J jnl
CoRR
Fabiana Fournier, Lior Limonad, Inna Skarbovsky
2023 J jnl
CoRR
Fabiana Fournier, Lior Limonad, Inna Skarbovsky, Yuval David
2023 conf
PMAI@IJCAI
Dirk Fahland, Fabiana Fournier, Lior Limonad, Inna Skarbovsky, Ava J. E. Swevels
2022 J jnl
IEEE Access
Gabriele Santin, Inna Skarbovsky, Fabiana Fournier, Bruno Lepri
2022 J jnl
CoRR
Gabriele Santin, Inna Skarbovsky, Fabiana Fournier, Bruno Lepri
2022 ch.
Technologies and Applications for Big Data Value
Óscar Lázaro, Jesús Alonso, Paulo Figueiras, Ruben Costa, Diogo Graça, Gisela García, Alessandro Canepa, Caterina Calefato, Marco Vallini, Fabiana Fournier, Nathan Hazout, Inna Skarbovsky, Athanasios Poulakidas, Konstantinos Sipsas
2022 conf
Business Process Management Workshops
Guy Amit, Fabiana Fournier, Lior Limonad, Inna Skarbovsky
2019 J jnl
IEEE Trans. Hum. Mach. Syst.
Natan Morar, Chris Baber, Faye McCabe, Sandra Dorothee Starke, Inna Skarbovsky, Alexander Artikis, Ivo Correai
2019 C conf
ICBC
Fabiana Fournier, Inna Skarbovsky
2017 conf
DEBS
Alexander Artikis, Nikos Katzouris, Ivo Correia, Chris Baber, Natan Morar, Inna Skarbovsky, Fabiana Fournier, Georgios Paliouras
2017 J jnl
CoRR
Alain Y. Kibangou, Alexander Artikis, Evangelos Michelioudakis, Georgios Paliouras, Marius Schmitt, John Lygeros, Chris Baber, Natan Morar, Fabiana Fournier, Inna Skarbovsky
2017 conf
DEBS
Ivo Correia, Alexander Artikis, Nikos Katzouris, Chris Baber, Natan Morar, Inna Skarbovsky, Fabiana Fournier, Georgios Paliouras
2016 conf
DEBS
Opher Etzion, Fabiana Fournier, Inna Skarbovsky, Barbara von Halle
2016 conf
DEBS
Ioannis Flouris, Vasiliki Manikaki, Nikos Giatrakos, Antonios Deligiannakis, Minos N. Garofalakis, Michael Mock, Sebastian Bothe, Inna Skarbovsky, Fabiana Fournier, Marko Stajcer, Tomislav Krizan, Jonathan Yom-Tov, Marijo Volarevic
2016 conf
SIGMOD Conference
Ioannis Flouris, Vasiliki Manikaki, Nikos Giatrakos, Antonios Deligiannakis, Minos N. Garofalakis, Michael Mock, Sebastian Bothe, Inna Skarbovsky, Fabiana Fournier, Marko Stajcer, Tomislav Krizan, Jonathan Yom-Tov, Taji Curin
2015 conf
EDBT/ICDT Workshops
Fabiana Fournier, Alexander Kofman, Inna Skarbovsky, Anastasios Skarlatidis
2015 conf
DEBS
Ivo Correia, Fabiana Fournier, Inna Skarbovsky
2011 conf
DEBS
Opher Etzion, Ella Rabinovich, Inna Skarbovsky
2010 conf
DEBS
Opher Etzion, Yonit Magid, Ella Rabinovich, Inna Skarbovsky, Nir Zolotorevsky
2008 conf
IEEE SCC (2)
Avivit Bercovici, Amit Fisher, Fabiana Fournier, Guy Rackham, Natalia Razinkov, Inna Skarbovsky
2005 conf
IEEE SCC
Asaf Adi, Dagan Gilat, Royi Ronen, Ron Rothblum, Guy Sharon, Inna Skarbovsky
2005 conf
RuleML
Dagan Gilat, Royi Ronen, Ron Rothblum, Guy Sharon, Inna Skarbovsky
2004 conf
WWW (Alternate Track Papers & Posters)
Asaf Adi, Opher Etzion, Dagan Gilat, Royi Ronen, Guy Sharon, Inna Skarbovsky
redb/extractors/pe_extractors/pe_extra_findings.py
← Index redb/extractors/pe_extractors/pe_extra_findings.py python
from hashlib import md5, sha1, sha256
import inspect
import pefile
from magika import Magika
import json
from datetime import datetime, timezone
from typing import List, Optional, Any
from asn1crypto import cms, pem, x509, core
from dataclasses import asdict

from redb.extractors.enum import Tag
from redb.extractors.pe_extractor import PEExtractor
from redb.models.dataclasses import PEExtraFinding, PECertificate


class PEExtraFindings(PEExtractor):

    def __init__(
        self,
        filepath,
        log,
        exporters=None,
        index_prefix=None,
        elastic_index=None,
        known_benign=False,
        known_malicious=False,
        pe=None,
    ):
        super().__init__(
            filepath,
            log,
            exporters,
            index_prefix,
            elastic_index,
            known_benign,
            known_malicious,
            pe,
        )
        self.elastic_index = self.index_prefix + "-pe_extrafindings"
        self.log.debug(inspect.currentframe().f_code.co_name)
        self.has_binary_overlay = False
        self.has_binary_resource = False
        self.findings = []

    def tag(self):
        return Tag.PE_EMBEDDED_EXTRAS.value

    """
    The section below triggers in case of a binary which is not signed.
    It trigger the scan for Certificates inside the binary, for example
    inside a resource or an overlay.
    """

    def _scan_for_certificates(self):
        """
        Comprehensive scan for certificate data in PE file.
        Returns list of tuples (location_description, certificate_data)
        """
        self.log.debug(inspect.currentframe().f_code.co_name)

        # 1. Scan resources for embedded PE files
        self._scan_resources_for_certificates()

        # 3. Scan overlay
        overlay_data = self.pe.get_overlay()
        if overlay_data:
            self._scan_overlay_for_certificates(overlay_data)


    def _extract_certificate_info(self, cert_data):
        self.log.debug(inspect.currentframe().f_code.co_name)

        certificate = x509.Certificate.load(cert_data)
        certificate_serial_number = format(certificate.serial_number, "x").upper()
        grouped_serial_number = ":".join(
            [
                certificate_serial_number[i : i + 2]
                for i in range(0, len(certificate_serial_number), 2)
            ]
        )
        return PECertificate(
            certificate_serial_number=grouped_serial_number,
            certificate_issuer=certificate.issuer.human_friendly,
            certificate_subject=certificate.subject.human_friendly,
            certificate_valid_from=certificate["tbs_certificate"]["validity"][
                "not_before"
            ].native.strftime("%Y-%m-%d %H:%M:%S"),
            certificate_valid_to=certificate["tbs_certificate"]["validity"][
                "not_after"
            ].native.strftime("%Y-%m-%d %H:%M:%S"),
            certificate_thumbprint=sha1(cert_data).hexdigest().upper(),
            certificate_algorithm=None,
        )

    def _extract_standard_certificates(self, pe):
        self.log.debug(inspect.currentframe().f_code.co_name)
        address = pe.OPTIONAL_HEADER.DATA_DIRECTORY[
            pefile.DIRECTORY_ENTRY["IMAGE_DIRECTORY_ENTRY_SECURITY"]
        ].VirtualAddress
        size = pe.OPTIONAL_HEADER.DATA_DIRECTORY[
            pefile.DIRECTORY_ENTRY["IMAGE_DIRECTORY_ENTRY_SECURITY"]
        ].Size

        addr_8 = address + 8
        signature_data = bytes(
            pe.write()[addr_8 : addr_8 + size]  # noqa E203
        )  # Ensure this is a bytes object
        if pem.detect(signature_data):
            signature_data = pem.unarmor(signature_data)

        content_info = cms.ContentInfo.load(signature_data)
        signed_data = content_info["content"]
        certificates = signed_data["certificates"]

        certificates_list = []
        for cert in certificates:
            cert_info = self._extract_certificate_info(cert.chosen.dump())
            certificates_list.append(cert_info)

        return certificates_list

    def _is_signed(self, pedata):
        address = pedata.OPTIONAL_HEADER.DATA_DIRECTORY[
            pefile.DIRECTORY_ENTRY["IMAGE_DIRECTORY_ENTRY_SECURITY"]
        ].VirtualAddress
        if address == 0:
            return False
        return True

    def _scan_resources_for_certificates(self):
        """Scan resources for embedded PE files and their certificates"""
        self.log.debug(inspect.currentframe().f_code.co_name)

        try:
            if hasattr(self.pe, "DIRECTORY_ENTRY_RESOURCE"):
                for entry in self.pe.DIRECTORY_ENTRY_RESOURCE.entries:
                    for resource in self._get_resource_data(entry):
                        # Check if resource data might be a PE file
                        resource_magika = Magika().identify_bytes(resource).output.label
                        if resource.startswith(b"MZ") or resource_magika == "pebin":
                            try:
                                embedded_pe = pefile.PE(data=resource)
                                if self._is_signed(embedded_pe):
                                    cert_data = self._extract_standard_certificates(
                                        embedded_pe
                                    )
                                    if cert_data:
                                        # certificates.append(
                                        #     (f"Resource_{entry.id}", cert_data)
                                        # )
                                        self.findings.append(
                                            PEExtraFinding(
                                                context=f"Resource ({sha256(resource).hexdigest()}) is a PE file. Found Code Signing Certificates information inside.",
                                                finding=cert_data,
                                            )
                                        )
                            except Exception as e:
                                self.log.warning(
                                    f"Resource {sha256(resource).hexdigest()} is not a valid PE: {e}"
                                )

                        # Direct certificate pattern matching in resource
                        cert_data = self._find_certificate_patterns(resource)
                        if cert_data:
                            # certificates.append((f"Resource_{entry.id}", cert_data))
                            self.findings.append(
                                PEExtraFinding(
                                    context=f"Resource ({sha256(resource).hexdigest()}) is a PE file.\
                                    Found Code Signing Certificates information inside.",
                                    finding=cert_data,
                                )
                            )

        except Exception as e:
            self.log.error(f"Failed to scan resources: {e}")


    def _scan_overlay_for_certificates(self, overlay_data) -> List[bytes]:
        """Scan PE overlay for certificate data"""
        self.log.debug(inspect.currentframe().f_code.co_name)
        try:
            # Check if overlay is a PE file
            overlay_magika = Magika().identify_bytes(overlay_data).output.ct_label
            if overlay_data.startswith(b"MZ") or overlay_magika == "pebin":
                try:
                    overlay_pe = pefile.PE(data=overlay_data)
                    if self._is_signed(overlay_pe):
                        cert_data = self._extract_standard_certificates(overlay_data)
                        if cert_data:
                            self.findings.append(
                                PEExtraFinding(
                                    context=f"Overlay data ({sha256(overlay_data).hexdigest()}) is a PE file.\
                                    Found Code Signing Certificates information inside.",
                                    finding=cert_data,
                                )
                            )
                except Exception as e:
                    self.log.warning(f"Error looking for certificates in Overlay: {e}")
                    pass

            # Direct certificate pattern matching in overlay
            cert_data = self._find_certificate_patterns(overlay_data)
            if cert_data:
                self.findings.append(
                    PEExtraFinding(
                        context=f"Overlay data ({sha256(overlay_data).hexdigest()}) is a PE file.\
                        Found Code Signing Certificates information inside.",
                        finding=cert_data,
                    )
                )
        except Exception as e:
            self.log.error(f"Failed to scan overlay: {e}")


    def _get_resource_data(self, entry) -> List[bytes]:
        """Recursively get resource data"""
        resources = []
        try:
            if hasattr(entry, "directory"):
                for subentry in entry.directory.entries:
                    resources.extend(self._get_resource_data(subentry))
            else:
                data = self.pe.get_data(
                    entry.data.struct.OffsetToData, entry.data.struct.Size
                )
                resources.append(data)
        except Exception as e:
            self.log.error(f"Failed to get resource data: {e}")
        return resources

    def _find_certificate_patterns(self, data: bytes) -> Optional[bytes]:
        """Find certificate patterns in binary data"""
        # PKCS#7 SignedData OID pattern
        PKCS7_PATTERN = b"\x06\x09\x2A\x86\x48\x86\xF7\x0D\x01\x07\x02"

        try:
            idx = data.find(PKCS7_PATTERN)
            if idx != -1:
                # Search backwards for ASN.1 sequence marker
                for i in range(idx, max(0, idx - 50), -1):
                    if data[i] == 0x30:  # ASN.1 SEQUENCE
                        try:
                            # Try to parse as X.509 certificate
                            cert = x509.load_der_x509_certificate(
                                data[i:], default_backend()
                            )
                            return data[i:]
                        except Exception:
                            continue
        except Exception as e:
            self.log.error(f"Failed to find certificate patterns: {e}")
        return None

    def prepare_export_data(self, exporter_type: str) -> Any:
        self.log.debug(inspect.currentframe().f_code.co_name)
        if exporter_type == "ElasticsearchExporter":
            return self.findings
        elif exporter_type == "ClickHouseExporter":
            current_time = datetime.now(timezone.utc)
            
            data = [
                [
                    self.sha256,
                    self.md5,
                    self.sha1,
                    finding.context,
                    [asdict(cert) for cert in finding.finding] if finding.finding else None,
                    current_time
                ]
                for finding in self.findings
            ]
            
            column_names = [
                'sha256', 'md5', 'sha1', 'context', 'finding', 'analysis_date'
            ]
            
            if not data:
                return None

            column_type_names = [
                'FixedString(64)', 'FixedString(32)', 'FixedString(40)',
                'String', 'Array(JSON)', 'DateTime64(3, \'UTC\')'
            ]

            return (data, column_names, column_type_names)

    def get_clickhouse_table(self) -> str:
        return "redb_extra_findings"

    def extract(self):
        try:
            self.log.debug(inspect.currentframe().f_code.co_name)
            self._scan_for_certificates()
            if len(self.findings) > 0:
                return self.findings
            return None
        except Exception as e:
            self.log.error(f"Error extracting PE Extras: {e}")
            return None