def manage_vulner_script(self,
test,
dupes,
script_element,
endpoint,
report_date=None):
for component_element in script_element.findall('table'):
component_cpe = CPE(component_element.attrib['key'])
for vuln in component_element.findall('table'):
# convert elements in dict
vuln_attributes = dict()
for elem in vuln.findall('elem'):
vuln_attributes[elem.attrib['key'].lower()] = elem.text
vuln_id = vuln_attributes['id']
description = "### Vulnerability\n\n"
description += "**ID**: `" + str(vuln_id) + "`\n"
description += "**CPE**: " + str(component_cpe) + "\n"
for attribute in vuln_attributes:
description += "**" + attribute + "**: `" + vuln_attributes[
attribute] + "`\n"
severity = self.convert_cvss_score(vuln_attributes['cvss'])
finding = Finding(
title=vuln_id,
test=test,
description=description,
severity=severity,
component_name=component_cpe.get_product()[0]
if len(component_cpe.get_product()) > 0 else '',
component_version=component_cpe.get_version()[0]
if len(component_cpe.get_version()) > 0 else '',
vuln_id_from_tool=vuln_id,
nb_occurences=1,
)
finding.unsaved_endpoints = [endpoint]
# manage if CVE is in metadata
if "type" in vuln_attributes and "cve" == vuln_attributes[
"type"]:
finding.cve = vuln_attributes["id"]
if report_date:
finding.date = report_date
dupe_key = finding.vuln_id_from_tool
if dupe_key in dupes:
find = dupes[dupe_key]
if description is not None:
find.description += "\n-----\n\n" + finding.description # fives '-' produces an horizontal line
find.unsaved_endpoints.extend(finding.unsaved_endpoints)
find.nb_occurences += finding.nb_occurences
else:
dupes[dupe_key] = finding
def manage_vulner_script(self, test, dupes, script_element, endpoint):
for component_element in script_element.findall('table'):
component_cpe = CPE(component_element.attrib['key'])
for vuln in component_element.findall('table'):
description = "### Vulnerability\n\n"
description += "**CPE**: " + str(component_cpe) + "\n"
vuln_attributes = dict()
for elem in vuln.findall('elem'):
vuln_attributes[elem.attrib['key'].lower()] = elem.text
description += "**" + elem.attrib['key'] + "**: " + elem.text + "\n"
cve = vuln_attributes['id']
severity = self.convert_cvss_score(vuln_attributes['cvss'])
dupe_key = cve
if dupe_key in dupes:
find = dupes[dupe_key]
if description is not None:
find.description += description
else:
find = Finding(title=cve,
cve=cve,
test=test,
description=description,
severity=severity,
mitigation="N/A",
impact="No impact provided",
component_name=component_cpe.get_product()[0] if len(component_cpe.get_product()) > 0 else '',
component_version=component_cpe.get_version()[0] if len(component_cpe.get_version()) > 0 else '',
)
find.unsaved_endpoints = list()
dupes[dupe_key] = find
find.unsaved_endpoints.append(endpoint)
async def load(self) -> None:
"""Load HassOS data."""
try:
if not self.sys_host.info.cpe:
raise NotImplementedError()
cpe = CPE(self.sys_host.info.cpe)
os_name = cpe.get_product()[0]
if os_name not in ("hassos", "haos"):
raise NotImplementedError()
except NotImplementedError:
_LOGGER.info("No Home Assistant Operating System found")
return
# Store meta data
self._available = True
self.sys_host.supported_features.cache_clear()
self._version = AwesomeVersion(cpe.get_version()[0])
self._board = cpe.get_target_hardware()[0]
self._os_name = cpe.get_product()[0]
await self.sys_dbus.rauc.update()
await self.datadisk.load()
_LOGGER.info(
"Detect Home Assistant Operating System %s / BootSlot %s",
self.version,
self.sys_dbus.rauc.boot_slot,
)
def config_cpe_match(self, cm):
if all("$.vulnerable", cm)[0]:
v = PLATFORM.VulnerableConfiguration
else:
v = PLATFORM.NotVulnerableConfiguration
subject = BNode()
cveStr = all("$.cpe23Uri", cm)[0]
self.triples(subject, v, [(PLATFORM.hasPlatform, cpeURI(cveStr))] + \
self.versionStartExcluding(cm) + self.versionStartIncluding(cm) + self.versionEndExcluding(cm) + self.versionEndIncluding(cm))
#print(cveStr)
c = CPE(cveStr)
if c.is_hardware():
self.g.add((cpeURI(cveStr), RDF.type, PLATFORM.HardwarePlatform))
elif c.is_application():
self.g.add(
(cpeURI(cveStr), RDF.type, PLATFORM.ApplicationPlatform))
elif c.is_operating_system():
self.g.add(
(cpeURI(cveStr), RDF.type, PLATFORM.OperatingSystemPlatform))
vendor = ""
for i in c.get_vendor():
self.g.add((cpeURI(cveStr), PLATFORM.vendor,
self.plEnt(i, "Vendor_", cls=PLATFORM.Vendor)))
vendor = i
for i in c.get_product():
self.g.add((cpeURI(cveStr), PLATFORM.product,
self.plEnt(i,
"Product_" + vendor + "_",
cls=PLATFORM.Product)))
for i in c.get_edition():
self.g.add((cpeURI(cveStr), PLATFORM.edition,
self.plEnt(i, "Edition_", cls=PLATFORM.Edition)))
for i in c.get_language():
self.g.add((cpeURI(cveStr), PLATFORM.language,
self.plEnt(i, "Language_", cls=PLATFORM.Language)))
for i in c.get_other():
self.g.add((cpeURI(cveStr), PLATFORM.other,
self.plEnt(i, "Other_", cls=PLATFORM.Other)))
for i in c.get_software_edition():
self.g.add((cpeURI(cveStr), PLATFORM.softwareEdition,
self.plEnt(i,
"SoftwareEdition_",
cls=PLATFORM.SoftwareEdition)))
for i in c.get_target_hardware():
self.g.add((cpeURI(cveStr), PLATFORM.targetHardware,
self.plEnt(i, "Hardware_", cls=CORE.Hardware)))
for i in c.get_target_software():
self.g.add((cpeURI(cveStr), PLATFORM.targetSoftware,
self.plEnt(i, "Software_", cls=CORE.Software)))
for i in c.get_update():
if not i == "-":
self.g.add((cpeURI(cveStr), PLATFORM.update, Literal(i)))
for i in c.get_version():
if not i == "-":
self.g.add((cpeURI(cveStr), PLATFORM.version, Literal(i)))
return subject
def get_description(self, address, name, cpe):
'''Define a description based on hostname and CPE'''
if name:
return name
else:
c = CPE(cpe[0], CPE.VERSION_2_3)
vendor = c.get_vendor()[0].upper()
if vendor in self.devs:
try:
client = SSHClient()
client.set_missing_host_key_policy(AutoAddPolicy())
client.connect(address,
username=self.devs[vendor]['USER'],
password=self.devs[vendor]['PASSWORD'])
stdin, stdout, stderr = client.exec_command(
self.devs[vendor]['COMMAND'])
return '{}:{}'.format(
vendor.lower(),
re.search(self.devs[vendor]['REGEX'],
str(stdout.read().decode('utf-8'))).group(
self.devs[vendor]['REGROUP']))
except (AuthenticationException, SSHException,
NoValidConnectionsError, TimeoutError,
ConnectionResetError):
pass
return '{}.{}.{}'.format(c.get_vendor()[0],
c.get_product()[0],
c.get_version()[0])
async def load(self) -> None:
"""Load HassOS data."""
try:
if not self.sys_host.info.cpe:
raise NotImplementedError()
cpe = CPE(self.sys_host.info.cpe)
if cpe.get_product()[0] != "hassos":
raise NotImplementedError()
except NotImplementedError:
_LOGGER.debug("Found no HassOS")
return
else:
self._available = True
# Store meta data
self._version = cpe.get_version()[0]
self._board = cpe.get_target_hardware()[0]
await self.sys_dbus.rauc.update()
_LOGGER.info("Detect HassOS %s / BootSlot %s", self.version,
self.sys_dbus.rauc.boot_slot)
with suppress(DockerAPIError):
await self.instance.attach(tag="latest")
def get_description(self, address, name, cpe):
'''Define a description based on hostname and CPE'''
if name:
return name
else:
c = CPE(cpe[0], CPE.VERSION_2_3)
vendor = c.get_vendor()[0]
if self.tacacs and vendor == 'cisco':
try:
client = SSHClient()
client.set_missing_host_key_policy(AutoAddPolicy())
client.connect(address,
username=self.tacacs['user'],
password=self.tacacs['password'])
stdin, stdout, stderr = client.exec_command(
self.tacacs['command'])
return '{}:{}'.format(
vendor.lower(),
re.search(self.tacacs['regex'],
str(stdout.read().decode('utf-8'))).group(
self.tacacs['regroup']))
except (AuthenticationException, SSHException,
NoValidConnectionsError, TimeoutError,
ConnectionResetError):
pass
return '{}.{}.{}'.format(c.get_vendor()[0],
c.get_product()[0],
c.get_version()[0])
def getinfo(filename):
if os.path.exists(filename + '.bin'):
print(filename, "exists -- taking in")
with open(filename + '.bin', 'rb') as reader:
return eval(reader.read())
print(filename, "does not exist, parsing")
f = open(filename)
tree = etree.parse(f)
f.close()
vulns = []
entry_nodes = tree.xpath(
'//prefix:entry',
namespaces={
'prefix': 'http://scap.nist.gov/schema/feed/vulnerability/2.0'
})
for entry in entry_nodes:
thisVuln = {}
thisVuln['id'] = entry.find(prefixed('vuln', 'cve-id')).text
vulnSoftware = entry.find(prefixed('vuln', 'vulnerable-software-list'))
if vulnSoftware is not None:
for v in vulnSoftware:
try:
myCPE = CPE(v.text)
except NotImplementedError:
print("Could not parse")
#logging.warning("Unable to parse CPE '%s'" % v.text)
else:
thisVuln['part'] = myCPE.get_part()[0]
thisVuln['vendor'] = myCPE.get_vendor()[0]
if 'linux' in thisVuln['vendor']:
thisVuln['vendor'] = 'linux'
thisVuln['product'] = myCPE.get_product()[0]
if 'linux' in thisVuln['product']:
thisVuln['vendor'] = 'linux'
thisVuln['version'] = myCPE.get_version()[0]
thisVuln['update'] = myCPE.get_update()[0]
thisVuln['edition'] = myCPE.get_edition()[0]
thisVuln['language'] = myCPE.get_language()[0]
cvss = entry.find(prefixed('vuln', 'cvss'))
if cvss is not None:
thisVuln['score'] = cvss.getchildren()[0].getchildren()[0].text
thisVuln['accessVector'] = cvss.getchildren()[0].getchildren(
)[1].text
thisVuln['accessComplexity'] = cvss.getchildren()[0].getchildren(
)[2].text
thisVuln['auth'] = cvss.getchildren()[0].getchildren()[3].text
thisVuln['impactConf'] = cvss.getchildren()[0].getchildren(
)[4].text
thisVuln['impactInt'] = cvss.getchildren()[0].getchildren()[5].text
thisVuln['impactAvail'] = cvss.getchildren()[0].getchildren(
)[6].text
vulns.append(thisVuln)
with open(filename + '.bin', 'wrb+') as myFile:
myFile.write(bytes(vulns))
def owasp_dependency_checker_task(source_code_location: str) -> List[Dict]:
"""
Run OWASP dependency-check and storage all vulnerabilities in
an unified format in Redis
"""
command = ['dependency-check --project "{loc}" --scan',
' "{loc}" -f "XML" -o {loc}',
' --enableExperimental']
# Install dependencies is VERY important
os.system("npm install")
# We use os.system instead of subprocess.call because because the OWASP
# Dependency check tool can take te CVE wordlist downloaded from the OS
# context. Using subprocess.call this is not possible. This implies
# that without the context no vulnerabilities was detected
os.system("".join(command).format(loc=source_code_location))
tree = ET.parse('{}/dependency-check-report.xml'. \
format(source_code_location))
root = tree.getroot()
SCHEME = "{https://jeremylong.github.io/DependencyCheck/dependency" \
"-check.1.3.xsd}"
results = []
for dependency in root.iterfind(".//{}dependency".format(SCHEME)):
# Get dep info
vulnerabilities = dependency.findall(".//{}vulnerability".format(SCHEME))
for vulnerability in vulnerabilities:
advisory = getattr(
vulnerability.find("{}name".format(SCHEME)),
"text", "")
severity = getattr(
vulnerability.find("{}severity".format(SCHEME)),
"text", "")
summary = getattr(
vulnerability.find("{}description".format(
SCHEME)),
"text", "")
for vulnerable_version in vulnerability.findall(
".//{}vulnerableSoftware/{}software["
"@allPreviousVersion='true']".format(
SCHEME, SCHEME)):
cpe = CPE(vulnerable_version.text)
library = cpe.get_product()[0]
version = cpe.get_version()[0]
results.append(dict(library=library,
version=version,
severity=severity,
summary=summary,
advisory=advisory))
return results
def __init__(self, cve_entry, thread: ThreadPool = None):
self.id = cve_entry.find(prefixed('vuln', 'cve-id')).text
self.vulnsoftware = cve_entry.find(
prefixed('vuln', 'vulnerable-software-list'))
self.part = []
self.vendor = []
self.version = []
self.update = []
self.edition = []
self.language = []
self.product = []
self.year_path = os.path.join("CVE_Detail", self.get_year())
self.cvssscore = None
self.accessVector = None
self.accessComplexity = None
self.auth = None
self.impactConf = None
self.impactInt = None
self.impactAvail = None
if self.vulnsoftware is not None:
for product in self.vulnsoftware:
try:
mycpe = CPE(product.text)
except NotImplementedError as e:
print(e)
else:
self.part.append(mycpe.get_part()[0])
self.vendor.append(mycpe.get_vendor()[0])
self.version.append(mycpe.get_version()[0])
self.update.append(mycpe.get_update()[0])
self.edition.append(mycpe.get_edition()[0])
self.language.append(mycpe.get_language()[0])
self.product.append(mycpe.get_product()[0])
cvss = cve_entry.find(prefixed('vuln', 'cvss'))
if cvss is not None:
self.cvssscore = cvss.getchildren()[0].getchildren()[0].text
self.accessVector = cvss.getchildren()[0].getchildren()[1].text
self.accessComplexity = cvss.getchildren()[0].getchildren()[2].text
self.auth = cvss.getchildren()[0].getchildren()[3].text
self.impactConf = cvss.getchildren()[0].getchildren()[4].text
self.impactInt = cvss.getchildren()[0].getchildren()[5].text
self.impactAvail = cvss.getchildren()[0].getchildren()[6].text
self.summery = cve_entry.find(prefixed('vuln', 'summary')).text
if thread != None:
thread.apply_async(self.get_from_pycvesearch)
async def load(self):
"""Load HassOS data."""
try:
assert self.sys_host.info.cpe is not None
cpe = CPE(self.sys_host.info.cpe)
assert cpe.get_product()[0] == 'hassos'
except (NotImplementedError, IndexError, AssertionError):
_LOGGER.info("Can't detect HassOS")
return
# Store meta data
self._available = True
self._version = cpe.get_version()[0]
self._board = cpe.get_target_hardware()[0]
_LOGGER.info("Detect HassOS %s on host system", self.version)
async def load(self):
"""Load HassOS data."""
try:
# Check needed host functions
assert self.sys_dbus.rauc.is_connected
assert self.sys_dbus.systemd.is_connected
assert self.sys_dbus.hostname.is_connected
assert self.sys_host.info.cpe is not None
cpe = CPE(self.sys_host.info.cpe)
assert cpe.get_product()[0] == 'hassos'
except (AssertionError, NotImplementedError):
_LOGGER.debug("Ignore HassOS")
return
# Store meta data
self._available = True
self._version = cpe.get_version()[0]
self._board = cpe.get_target_hardware()[0]
_LOGGER.info("Detect HassOS %s on host system", self.version)
async def load(self) -> None:
"""Load HassOS data."""
try:
# Check needed host functions
assert self.sys_dbus.rauc.is_connected
assert self.sys_dbus.systemd.is_connected
assert self.sys_dbus.hostname.is_connected
assert self.sys_host.info.cpe is not None
cpe = CPE(self.sys_host.info.cpe)
assert cpe.get_product()[0] == "hassos"
except (AssertionError, NotImplementedError):
_LOGGER.debug("Found no HassOS")
return
# Store meta data
self._available = True
self._version = cpe.get_version()[0]
self._board = cpe.get_target_hardware()[0]
_LOGGER.info("Detect HassOS %s on host system", self.version)
with suppress(DockerAPIError):
await self.instance.attach()
async def load(self) -> None:
"""Load HassOS data."""
try:
if not self.sys_host.info.cpe:
raise NotImplementedError()
cpe = CPE(self.sys_host.info.cpe)
if cpe.get_product()[0] != "hassos":
raise NotImplementedError()
except NotImplementedError:
_LOGGER.info("No Home Assistant Operating System found")
return
else:
self._available = True
# Store meta data
self._version = cpe.get_version()[0]
self._board = cpe.get_target_hardware()[0]
await self.sys_dbus.rauc.update()
_LOGGER.info("Detect HassOS %s / BootSlot %s", self.version,
self.sys_dbus.rauc.boot_slot)
def sync_es_index(cvesearch_url, esd_url, namelen):
"""
synchronize vulnsearch esd index with cvesearch data for all cpe notes in storage
"""
esclient = Elasticsearch([esd_url])
indexer = BulkIndexer(esclient)
# create new index. there can be new or updated services or cpe notes of
# existing services, also there might be new or updated cves for cpes the
# easiest way to handle such complex situation is to create new index and
# update alias used by the vulnsearch elk ui objects
current_index = f'{ES_INDEX}-{time()}'
for note in windowed_query(Note.query.filter(Note.xtype == 'cpe'), Note.id):
for icpe in json.loads(note.data):
try:
parsed_cpe = CPE(icpe)
except Exception: # pylint: disable=broad-except ; library does not provide own core exception class
current_app.logger.warning(f'invalid cpe, note_id:{note.id} {icpe}')
continue
if not parsed_cpe.get_version()[0]:
continue
for cve in cvefor(icpe, cvesearch_url):
data_id, data = vulndata(note, parsed_cpe, cve, namelen)
indexer.index(current_index, data_id, data)
indexer.flush()
# update alias and prune old indexes
update_managed_indices(esclient, current_index)
# print cache stats
current_app.logger.debug(f'cvefor cache: {cvefor.cache_info()}') # pylint: disable=no-value-for-parameter ; lru decorator side-effect
async def load(self) -> None:
"""Load OppOS data."""
try:
if not self.sys_host.info.cpe:
raise NotImplementedError()
cpe = CPE(self.sys_host.info.cpe)
if cpe.get_product()[0] != "oppos":
raise NotImplementedError()
except NotImplementedError:
_LOGGER.info("No Open Peer Power Operating System found")
return
else:
self._available = True
self.sys_host.supported_features.cache_clear()
# Store meta data
self._version = AwesomeVersion(cpe.get_version()[0])
self._board = cpe.get_target_hardware()[0]
await self.sys_dbus.rauc.update()
_LOGGER.info("Detect OppOS %s / BootSlot %s", self.version,
self.sys_dbus.rauc.boot_slot)
def populate_CVE(root):
cve_data = []
vuln_data = []
for entry in root:
cve_id = entry.find(prefixed('vuln', 'cve-id')).text
cve_id = int(re.sub("[^0-9]", "", cve_id))
pubdate = entry.find(prefixed('vuln', 'published-datetime')).text
moddate = entry.find(prefixed('vuln', 'last-modified-datetime')).text
summary = entry.find(prefixed('vuln', 'summary')).text
pubdate = parser.parse(pubdate)
moddate = parser.parse(moddate)
vulnSoftware = entry.find(prefixed('vuln', 'vulnerable-software-list'))
vulnList = []
unableToParse=0
if vulnSoftware is not None:
for v in vulnSoftware:
try:
myCPE = CPE(v.text)
except NotImplementedError:
unableToParse+=1
#logging.warning("Unable to parse CPE '%s'" % v.text)
else:
part = myCPE.get_part()[0]
vendor = myCPE.get_vendor()[0]
product = myCPE.get_product()[0]
version = myCPE.get_version()[0]
update = myCPE.get_update()[0]
edition = myCPE.get_edition()[0]
language = myCPE.get_language()[0]
derpa = {"part" : part, "vendor":vendor, "product":product, "version":version, "update":update, "edition":edition, "language":language, "cve":cve_id}
vuln_data.append(derpa)
if unableToParse>0:
logging.warning("Could not parse %d lines from file." % unableToParse)
vuln = entry.find(prefixed('vuln','cvss'))
#metrics = vuln.find(prefixed('cvss','base_metrics'))
if vuln is not None:
score = vuln.getchildren()[0].getchildren()[0].text
accessVector = vuln.getchildren()[0].getchildren()[1].text
accessComplexity = vuln.getchildren()[0].getchildren()[2].text
auth = vuln.getchildren()[0].getchildren()[3].text
impactConf = vuln.getchildren()[0].getchildren()[4].text
impactInt = vuln.getchildren()[0].getchildren()[5].text
impactAvail = vuln.getchildren()[0].getchildren()[6].text
if "DO NOT USE THIS CANDIDATE NUMBER" not in summary:
data = {
"cve":cve_id,
"pubdate":pubdate,
"moddate":moddate,
"summary":summary,
"score":score,
"accessVector":accessVector,
"accessComp":accessComplexity,
"auth":auth,
"impactConf": impactConf,
"impactInt": impactInt,
"impactAvail": impactAvail
}
cve_data.append(data)
tables['CVEs'].insert().execute(cve_data)
tables['VulnSoftware'].insert().execute(vuln_data)
def get_findings(self, filename, test: Test):
content = filename.read()
if type(content) is bytes:
content = content.decode('utf-8')
csv.field_size_limit(int(sys.maxsize / 10)) # the request/resp are big
reader = csv.DictReader(io.StringIO(content))
dupes = dict()
for row in reader:
# manage severity from two possible columns 'Severity' and 'Risk'
severity = 'Info'
if 'Severity' in row:
severity = self._convert_severity(row.get('Severity'))
elif 'Risk' in row:
severity = self._convert_severity(row.get('Risk'))
# manage title from two possible columns 'Nme' and 'Plugin Name'
title = row.get('Name')
if title is None and 'Plugin Name' in row:
title = row.get('Plugin Name')
# special case to skip empty titles
if not title:
continue
description = row.get('Synopsis')
mitigation = str(row.get('Solution'))
impact = row.get('Description', 'N/A')
references = row.get('See Also', 'N/A')
dupe_key = severity + title + row.get('Host', 'No host') + str(
row.get('Port', 'No port')) + row.get('Synopsis',
'No synopsis')
detected_cve = self._format_cve(str(row.get('CVE')))
cve = None
if detected_cve:
# FIXME support more than one CVE in Nessus CSV parser
cve = detected_cve[0]
if len(detected_cve) > 1:
LOGGER.warning(
"more than one CVE for a finding. NOT supported by Nessus CSV parser"
)
if dupe_key in dupes:
find = dupes[dupe_key]
if 'Plugin Output' in row:
find.description += row.get('Plugin Output')
else:
if 'Plugin Output' in row:
description = description + str(row.get('Plugin Output'))
find = Finding(title=title,
test=test,
cve=cve,
description=description,
severity=severity,
mitigation=mitigation,
impact=impact,
references=references)
# manage CVSS vector (only v3.x for now)
if 'CVSS V3 Vector' in row and '' != row.get('CVSS V3 Vector'):
find.cvssv3 = CVSS3('CVSS:3.0/' +
str(row.get('CVSS V3 Vector'))
).clean_vector(output_prefix=False)
# manage CPE data
detected_cpe = self._format_cpe(str(row.get('CPE')))
if detected_cpe:
# FIXME support more than one CPE in Nessus CSV parser
if len(detected_cpe) > 1:
LOGGER.warning(
"more than one CPE for a finding. NOT supported by Nessus CSV parser"
)
cpe_decoded = CPE(detected_cpe[0])
find.component_name = cpe_decoded.get_product()[0] if len(
cpe_decoded.get_product()) > 0 else None
find.component_version = cpe_decoded.get_version(
)[0] if len(cpe_decoded.get_version()) > 0 else None
find.unsaved_endpoints = list()
dupes[dupe_key] = find
# manage endpoints
endpoint = Endpoint(host='localhost')
if 'Host' in row:
endpoint.host = row.get('Host')
elif 'IP Address' in row:
endpoint.host = row.get('IP Address')
endpoint.port = row.get('Port')
if 'Protocol' in row:
endpoint.protocol = row.get('Protocol').lower()
find.unsaved_endpoints.append(endpoint)
return list(dupes.values())
def get_component_name_and_version_from_dependency(self, dependency,
related_dependency,
namespace):
component_name, component_version = None, None
# big try catch to avoid crashint the parser on some unexpected stuff
try:
identifiers_node = dependency.find(namespace + 'identifiers')
if identifiers_node:
# <identifiers>
# <identifier type="cpe" confidence="HIGHEST">
# <name>cpe:/a:apache:xalan-java:2.7.1</name>
# <url>https://web.nvd.nist.gov/view/vuln/search-results?adv_search=true&cves=on&cpe_version=cpe%3A%2Fa%3Aapache%3Axalan-java%3A2.7.1</url>
# </identifier>
# <identifier type="maven" confidence="HIGHEST">
# <name>xalan:serializer:2.7.1</name>
# <url>https://search.maven.org/remotecontent?filepath=xalan/serializer/2.7.1/serializer-2.7.1.jar</url>
# </identifier>
# </identifiers>
# newly found in v6.0.0
# <identifiers>
# <package confidence="HIGH">
# <id>pkg:maven/nl.isaac.client.offerservice/[email protected]</id>
# <url>https://ossindex.sonatype.org/component/pkg:maven/nl.isaac.client.offerservice/[email protected]</url>
# </package>
# </identifiers>
# <identifiers>
# <package confidence="HIGHEST">
# <id>pkg:npm/[email protected]</id>
# <url>https://ossindex.sonatype.org/component/pkg:npm/[email protected]</url>
# </package>
# </identifiers>
package_node = identifiers_node.find('.//' + namespace +
'package')
if package_node:
logger.debug(
'package string: ' +
self.get_field_value(package_node, 'id', namespace))
id = self.get_field_value(package_node, 'id', namespace)
purl = PackageURL.from_string(id)
purl_parts = purl.to_dict()
component_name = purl_parts[
'namespace'] + ':' if purl_parts['namespace'] and len(
purl_parts['namespace']) > 0 else ''
component_name += purl_parts['name'] if purl_parts[
'name'] and len(purl_parts['name']) > 0 else ''
component_name = component_name if component_name else None
component_version = purl_parts['version'] if purl_parts[
'version'] and len(purl_parts['version']) > 0 else ''
return component_name, component_version
cpe_node = identifiers_node.find('.//' + namespace +
'identifier[@type="cpe"]')
if cpe_node:
# logger.debug('cpe string: ' + self.get_field_value(cpe_node, 'name'))
cpe = CPE(self.get_field_value(cpe_node, 'name'))
component_name = cpe.get_vendor()[0] + ':' if len(
cpe.get_vendor()) > 0 else ''
component_name += cpe.get_product()[0] if len(
cpe.get_product()) > 0 else ''
component_name = component_name if component_name else None
component_version = cpe.get_version()[0] if len(
cpe.get_version()) > 0 else None
# logger.debug('get_edition: ' + str(cpe.get_edition()))
# logger.debug('get_language: ' + str(cpe.get_language()))
# logger.debug('get_part: ' + str(cpe.get_part()))
# logger.debug('get_software_edition: ' + str(cpe.get_software_edition()))
# logger.debug('get_target_hardware: ' + str(cpe.get_target_hardware()))
# logger.debug('get_target_software: ' + str(cpe.get_target_software()))
# logger.debug('get_vendor: ' + str(cpe.get_vendor()))
# logger.debug('get_update: ' + str(cpe.get_update()))
return component_name, component_version
maven_node = identifiers_node.find('.//' + namespace +
'identifier[@type="maven"]')
if maven_node:
# logger.debug('maven_string: ' + self.get_field_value(maven_node, 'name'))
maven_parts = self.get_field_value(maven_node, 'name',
namespace).split(':')
# logger.debug('maven_parts:' + str(maven_parts))
if len(maven_parts) == 3:
component_name = maven_parts[0] + ':' + maven_parts[1]
component_version = maven_parts[2]
return component_name, component_version
# TODO
# include identifiers in description?
# <identifiers>
# <package confidence="HIGH">
# <id>pkg:maven/org.dom4j/[email protected]</id>
# <url>https://ossindex.sonatype.org/component/pkg:maven/org.dom4j/[email protected]</url>
# </package>
# <vulnerabilityIds confidence="HIGHEST">
# <id>cpe:2.3:a:dom4j_project:dom4j:2.1.1.hat-00001:*:*:*:*:*:*:*</id>
# <url>https://nvd.nist.gov/vuln/search/results?form_type=Advanced&results_type=overview&search_type=all&cpe_vendor=cpe%3A%2F%3Adom4j_project&cpe_product=cpe%3A%2F%3Adom4j_project%3Adom4j&cpe_version=cpe%3A%2F%3Adom4j_project%3Adom4j%3A2.1.1.hat-00001</url>
# </vulnerabilityIds>
# TODO what happens when there multiple evidencecollectednodes with product or version as type?
evidence_collected_node = dependency.find(namespace +
'evidenceCollected')
if evidence_collected_node:
# <evidenceCollected>
# <evidence type="product" confidence="HIGH">
# <source>file</source>
# <name>name</name>
# <value>jquery</value>
# </evidence>
# <evidence type="version" confidence="HIGH">
# <source>file</source>
# <name>version</name>
# <value>3.1.1</value>
# </evidence>'
# will find the first product and version node. if there are multiple it may not pick the best
# since 6.0.0 howoever it seems like there's always a packageurl above so not sure if we need the effort to
# implement more logic here
product_node = evidence_collected_node.find(
'.//' + namespace + 'evidence[@type="product"]')
if product_node:
component_name = self.get_field_value(
product_node, 'value', namespace)
version_node = evidence_collected_node.find(
'.//' + namespace + 'evidence[@type="version"]')
if version_node:
component_version = self.get_field_value(
version_node, 'value', namespace)
return component_name, component_version
except:
logger.exception(
'error parsing component_name and component_version')
logger.debug(
'dependency: %s',
ElementTree.tostring(dependency, encoding='utf8',
method='xml'))
return component_name, component_version
def get_component_name_and_version_from_dependency(self, dependency,
related_dependency,
namespace):
identifiers_node = dependency.find(namespace + 'identifiers')
if identifiers_node:
# analyzing identifier from the more generic to
package_node = identifiers_node.find('.//' + namespace + 'package')
if package_node:
id = package_node.findtext(f'{namespace}id')
purl = PackageURL.from_string(id)
purl_parts = purl.to_dict()
component_name = purl_parts['namespace'] + ':' if purl_parts[
'namespace'] and len(purl_parts['namespace']) > 0 else ''
component_name += purl_parts['name'] if purl_parts[
'name'] and len(purl_parts['name']) > 0 else ''
component_name = component_name if component_name else None
component_version = purl_parts['version'] if purl_parts[
'version'] and len(purl_parts['version']) > 0 else ''
return component_name, component_version
# vulnerabilityIds_node = identifiers_node.find('.//' + namespace + 'vulnerabilityIds')
# if vulnerabilityIds_node:
# id = vulnerabilityIds_node.findtext(f'{namespace}id')
# cpe = CPE(id)
# component_name = cpe.get_vendor()[0] + ':' if len(cpe.get_vendor()) > 0 else ''
# component_name += cpe.get_product()[0] if len(cpe.get_product()) > 0 else ''
# component_name = component_name if component_name else None
# component_version = cpe.get_version()[0] if len(cpe.get_version()) > 0 else None
# return component_name, component_version
cpe_node = identifiers_node.find('.//' + namespace +
'identifier[@type="cpe"]')
if cpe_node:
id = cpe_node.findtext(f'{namespace}name')
cpe = CPE(id)
component_name = cpe.get_vendor()[0] + ':' if len(
cpe.get_vendor()) > 0 else ''
component_name += cpe.get_product()[0] if len(
cpe.get_product()) > 0 else ''
component_name = component_name if component_name else None
component_version = cpe.get_version()[0] if len(
cpe.get_version()) > 0 else None
return component_name, component_version
maven_node = identifiers_node.find('.//' + namespace +
'identifier[@type="maven"]')
if maven_node:
maven_parts = maven_node.findtext(f'{namespace}name').split(
':')
# logger.debug('maven_parts:' + str(maven_parts))
if len(maven_parts) == 3:
component_name = maven_parts[0] + ':' + maven_parts[1]
component_version = maven_parts[2]
return component_name, component_version
# TODO what happens when there multiple evidencecollectednodes with product or version as type?
evidence_collected_node = dependency.find(namespace +
'evidenceCollected')
if evidence_collected_node:
# <evidenceCollected>
# <evidence type="product" confidence="HIGH">
# <source>file</source>
# <name>name</name>
# <value>jquery</value>
# </evidence>
# <evidence type="version" confidence="HIGH">
# <source>file</source>
# <name>version</name>
# <value>3.1.1</value>
# </evidence>'
# will find the first product and version node. if there are multiple it may not pick the best
# since 6.0.0 howoever it seems like there's always a packageurl above so not sure if we need the effort to
# implement more logic here
product_node = evidence_collected_node.find(
'.//' + namespace + 'evidence[@type="product"]')
if product_node:
component_name = product_node.findtext(f'{namespace}value')
version_node = evidence_collected_node.find(
'.//' + namespace + 'evidence[@type="version"]')
if version_node:
component_version = version_node.findtext(
f'{namespace}value')
return component_name, component_version
return None, None
def get_cpe_df(self, debug=False):
"""Get the list of CPE names for the vulnerability.
"""
type_list = []
part_list = []
vendor_list = []
product_list = []
version_list = []
update_list = []
edition_list = []
language_list = []
sw_edition_list = []
target_sw_list = []
target_hw_list = []
other_list = []
published_datetime_list = []
for cpe_entry in self.cpe_list:
#if(debug):
#print(cpe_entry)
try:
cp = CPE(cpe_entry)
if(cp.is_hardware()):
type_list.append("HW")
elif(cp.is_operating_system()):
type_list.append("OS")
elif(cp.is_application()):
type_list.append("APP")
else:
type_list.append("UNDEFINED")
part_list.append(list_to_string(cp.get_part()))
vendor_list.append(list_to_string(cp.get_vendor()))
product_list.append(list_to_string(cp.get_product()))
version_list.append(list_to_string(cp.get_version()))
update_list.append(list_to_string(cp.get_update()))
edition_list.append(list_to_string(cp.get_edition()))
language_list.append(list_to_string(cp.get_language()))
sw_edition_list.append(list_to_string(cp.get_software_edition()))
target_sw_list.append(list_to_string(cp.get_target_software()))
target_hw_list.append(list_to_string(cp.get_target_hardware()))
other_list.append(list_to_string(cp.get_other()))
published_datetime_list.append(self.published_datetime)
except Exception as inst:
print(inst)
data = pd.DataFrame()
data['type'] = type_list
data['part'] = part_list
data['vendor'] = vendor_list
data['product'] = product_list
data['version'] = version_list
data['update'] = update_list
data['edition'] = edition_list
data['language'] = language_list
data['sw_edition'] = sw_edition_list
data['target_sw'] = target_sw_list
data['target_hw'] = target_hw_list
data['other'] = other_list
data['published_datetime'] = published_datetime_list
return data
class Service(object):
"""
Represents service/application/operating system. Contains basic information: name, version
"""
_CPE_SPECIAL = r"\!|\"|\;|\#|\$|\%|\&|\'|\(|\)|\+|\,|\/|\:|\<|\=|\>|\@|\[|\]|\^|\`|\{|\||\}|\~|\-"
_ESCAPE_CPE = re.compile(_CPE_SPECIAL)
_UNESCAPE_CPE = re.compile(r"(\\({0}))".format(_CPE_SPECIAL))
def __init__(self, name=None, version=None, cpe=None):
self.name = name
self.version = version
self._cpe = None
self.cpe = cpe
@property
def cpe(self) -> CPE:
"""
CPE representation of service
"""
return self._cpe
@cpe.setter
def cpe(self, value):
if value:
self._cpe = CPE(value)
@property
def cpe_vendor(self) -> str:
"""
Get vendor name based on CPE
"""
if isinstance(self._cpe, CPE):
return self._unescape_cpe(" ".join(self._cpe.get_vendor()))
@property
def name_with_version(self) -> str:
"""
Service name with version included
"""
if self.version is None or self.name is None:
return None
return "{name} {version}".format(name=self.name, version=self.version)
@property
def cpe_product(self) -> str:
"""
Get product name based on CPE
"""
if isinstance(self._cpe, CPE):
return self._unescape_cpe(" ".join(self._cpe.get_product()))
@property
def cpe_version(self) -> str:
"""
Get product name based on CPE
"""
if isinstance(self._cpe, CPE):
return self._unescape_cpe(" ".join(self._cpe.get_version()))
def __str__(self):
return "{name} {version}".format(name=self.name or '', version=self.version or '').strip()
def copy(self) -> 'Service':
"""
Make copy of service
"""
return_value = Service(name=self.name, version=self.version)
return_value._cpe = self._cpe
return return_value
@classmethod
def _escape_cpe(cls, text: str) -> str:
"""
Special characters should be escaped before building CPE string
"""
text = text.lower()
def _replace(txt):
return r"\{0}".format(txt.group())
if " " in text:
raise ValueError("{0}: Space is not allowed in CPE string".format(text))
return cls._ESCAPE_CPE.sub(_replace, text)
@classmethod
def _unescape_cpe(cls, text: str) -> str:
text = text.lower()
def _replace(txt):
return txt.group()[1]
return cls._UNESCAPE_CPE.sub(_replace, text)
@classmethod
def validate_cpe_arguments(cls, vendor: str, product: str, version: str) -> (str, str, str):
"""
Validate cpe arguments, and fix as much as possible
"""
if " " in version:
if product.lower() == "ios":
version = version.split(" ")[0].strip(",")
if vendor == "*":
if product.lower() == "ios":
vendor = "cisco"
return cls._escape_cpe(vendor), cls._escape_cpe(product), cls._escape_cpe(version)
@classmethod
def build_cpe(cls, part: 'CPEType', vendor: str = '*', product: str = '*', version: str = '*') -> str:
"""
Build cpe 2.3 string base on vendor, product, version and part
"""
vendor, product, version = cls.validate_cpe_arguments(vendor=vendor, product=product, version=version)
return "cpe:2.3:{part}:{vendor}:{product}:{version}:*:*:*:*:*:*:*".format(part=str(part.value), vendor=vendor,
product=product, version=version)
def extract_vendor_product_version(cpe_str):
"""Extract vendor and product from NVD cve entry."""
cpe_ = CPE(cpe_str)
return cpe_.get_vendor()[0], cpe_.get_product()[0], cpe_.get_version()[0]
def get_item(vuln, test):
finding = Finding(
test=test,
unique_id_from_tool=vuln["id"],
nb_occurences=1,
)
# Defining variables
location = vuln["location"]
# Endpoint
# using url
if "url" in location and location["url"] and location["url"] != "None":
endpoint = Endpoint.from_uri(location["url"])
# fallback to using old way of creating endpoints
elif "domain" in location and location["domain"] and location["domain"] != "None":
endpoint = Endpoint(host=str(location["domain"]))
else: # no domain, use ip instead
if "ip" in location and location["ip"] and location["ip"] != "None":
endpoint = Endpoint(host=str(location["ip"]))
# check for protocol
if (
"applicationProtocol" in location and
location["applicationProtocol"] and
location["applicationProtocol"] != "None"
):
endpoint.protocol = location["applicationProtocol"]
# check for port
if (
"port" in location and
location["port"] in location and
location["port"] != "None"
):
endpoint.port = location["port"]
finding.unsaved_endpoints = [endpoint] # assigning endpoint
# Title
finding.title = vuln["name"]
# Description + CVEs
description = vuln["classification"]
cves = "no match"
if "CVE-NO-MATCH" not in vuln["kb"]["cves"]:
finding.cve = vuln["kb"]["cves"][0]
cves = ""
for cve in vuln["kb"]["cves"]:
cves += f"{cve}, "
cves = cves[: len(cves) - 2] # removing the comma and the blank space
finding.description = description + "; CVEs: " + cves
finding.severity = vuln["severity"].title()
# Date
date_str = vuln["createdOn"]
date_str = date_str[: len(date_str) - 3] + date_str[-2:]
finding.date = datetime.strptime(date_str, "%Y-%m-%dT%H:%M:%S.%f%z")
# Component Name and Version
if (
"applicationCpe" in location and
location["applicationCpe"] and
location["applicationCpe"] != "None"
):
cpe = CPE(location["applicationCpe"])
component_name = cpe.get_vendor()[0] + ":" if len(
cpe.get_vendor()) > 0 else ""
component_name += cpe.get_product()[0] if len(
cpe.get_product()) > 0 else ""
finding.component_name = component_name if component_name else None
finding.component_version = (
cpe.get_version()[0] if len(cpe.get_version()) > 0 else None
)
return finding
def populate_CVE(root):
cve_data = []
vuln_data = []
for entry in root:
cve_id = entry.find(prefixed("vuln", "cve-id")).text
cve_id = int(re.sub("[^0-9]", "", cve_id))
pubdate = entry.find(prefixed("vuln", "published-datetime")).text
moddate = entry.find(prefixed("vuln", "last-modified-datetime")).text
summary = entry.find(prefixed("vuln", "summary")).text
pubdate = parser.parse(pubdate)
moddate = parser.parse(moddate)
vulnSoftware = entry.find(prefixed("vuln", "vulnerable-software-list"))
vulnList = []
unableToParse = 0
if vulnSoftware is not None:
for v in vulnSoftware:
try:
myCPE = CPE(v.text)
except NotImplementedError:
unableToParse += 1
# logging.warning("Unable to parse CPE '%s'" % v.text)
else:
part = myCPE.get_part()[0]
vendor = myCPE.get_vendor()[0]
product = myCPE.get_product()[0]
version = myCPE.get_version()[0]
update = myCPE.get_update()[0]
edition = myCPE.get_edition()[0]
language = myCPE.get_language()[0]
derpa = {
"part": part,
"vendor": vendor,
"product": product,
"version": version,
"update": update,
"edition": edition,
"language": language,
"cve": cve_id,
}
vuln_data.append(derpa)
if unableToParse > 0:
logging.warning("Could not parse %d lines from file." % unableToParse)
vuln = entry.find(prefixed("vuln", "cvss"))
# metrics = vuln.find(prefixed('cvss','base_metrics'))
if vuln is not None:
score = vuln.getchildren()[0].getchildren()[0].text
accessVector = vuln.getchildren()[0].getchildren()[1].text
accessComplexity = vuln.getchildren()[0].getchildren()[2].text
auth = vuln.getchildren()[0].getchildren()[3].text
impactConf = vuln.getchildren()[0].getchildren()[4].text
impactInt = vuln.getchildren()[0].getchildren()[5].text
impactAvail = vuln.getchildren()[0].getchildren()[6].text
if "DO NOT USE THIS CANDIDATE NUMBER" not in summary:
data = {
"cve": cve_id,
"pubdate": pubdate,
"moddate": moddate,
"summary": summary,
"score": score,
"accessVector": accessVector,
"accessComp": accessComplexity,
"auth": auth,
"impactConf": impactConf,
"impactInt": impactInt,
"impactAvail": impactAvail,
}
cve_data.append(data)
tables["CVEs"].insert().execute(cve_data)
tables["VulnSoftware"].insert().execute(vuln_data)
