def parse_filename(self, filename):
"""
returns package_name/product_name from package_filename of types .rpm, .deb, etc.
Example: package_filename = openssh-client_8.4p1-5ubuntu1_amd64.deb
here, package_name = openssh-client
"""
# resolving directory names
if sys.platform == "win32":
filename = filename.split("\\")[-1]
else:
filename = filename.split("/")[-1]
# if extractable, then parsing for different types of files accordingly
if self.extractor.can_extract(filename):
if filename.endswith(".tar.xz"):
product_name = filename.rsplit("-", 3)[0]
version_number = filename.rsplit("-", 3)[1]
# example: libarchive-3.5.1-1-aarch64.pkg.tar.xz
elif filename.endswith(".deb") or filename.endswith(".ipk"):
product_name = filename.rsplit("_")[0]
version_number = filename.rsplit("_")[1]
# example: varnish_6.4.0-3_amd64.deb
else:
product_name = filename.rsplit("-", 2)[0]
version_number = filename.rsplit("-", 2)[1]
LOGGER.debug(
f"Parsing file '{self.filename}': Results: product_name='{product_name}', version_number='{version_number}'"
)
return product_name, version_number
else:
# raise error for unknown archive types
with ErrorHandler(mode=ErrorMode.NoTrace, logger=LOGGER):
raise UnknownArchiveType(filename)
def parse_spdx_rdf(self, sbom_file: str) -> List[List[str]]:
"""parses SPDX RDF BOM file extracting package name and version"""
with open(sbom_file) as f:
lines = f.readlines()
modules: List[List[str]] = []
package = ""
for line in lines:
try:
if line.strip().startswith("<spdx:name>"):
stripped_line = line.strip().rstrip("\n")
package_match = re.search(
"<spdx:name>(.+?)</spdx:name>", stripped_line
)
if not package_match:
raise KeyError(f"Could not find package in {stripped_line}")
package = package_match.group(1)
version = None
elif line.strip().startswith("<spdx:versionInfo>"):
stripped_line = line.strip().rstrip("\n")
version_match = re.search(
"<spdx:versionInfo>(.+?)</spdx:versionInfo>", stripped_line
)
if not version_match:
raise KeyError(f"Could not find version in {stripped_line}")
version = version_match.group(1)
modules.append([package, version])
except KeyError as e:
LOGGER.debug(e, exc_info=True)
return modules
def parse_spdx_xml(self, sbom_file: str) -> List[List[str]]:
"""parses SPDX XML BOM file extracting package name and version"""
# XML is experimental in SPDX 2.2
tree = ET.parse(sbom_file)
# Find root element
root = tree.getroot()
# Extract schema
schema = root.tag[: root.tag.find("}") + 1]
modules: List[List[str]] = []
for component in root.findall(schema + "packages"):
try:
package_match = component.find(schema + "name")
if package_match is None:
raise KeyError(f"Could not find package in {component}")
package = package_match.text
if package is None:
raise KeyError(f"Could not find package in {component}")
version_match = component.find(schema + "versionInfo")
if version_match is None:
raise KeyError(f"Could not find version in {component}")
version = version_match.text
if version is None:
raise KeyError(f"Could not find version in {component}")
modules.append([package, version])
except KeyError as e:
LOGGER.debug(e, exc_info=True)
return modules
def search_version_string(self, matched_list):
"""finds version strings from matched list"""
# TODO: add multiline string finding
pattern1 = rf"{self.product_name}(.*){self.version_number}"
# ^ this does not work for debian packages
# pattern2 = rf"{self.product_name}(.*)([0-9]+[.-][0-9]+([.-][0-9]+)?)"
# this matches patterns like:
# product1.2.3
# product 1.2.3
# product-1.2.3
# product.1.2.3
# product version 1.2.3
# product v1.2.3(1)
version_strings = [
i for i in matched_list
if re.search(pattern1, i, re.IGNORECASE) if not i.endswith(
".debug") # removes .debug, so, this does not gets printed
]
LOGGER.debug(f"found version-string matches = {version_strings}"
) # TODO: regex highlight in these matched strings?
return version_strings
def get_data(self, cve_number: str, product: str):
try:
full_query = f"{RH_CVE_API}/{cve_number}.json" # static https url above
response = request.urlopen(full_query).read().decode(
"utf-8") # nosec
return loads(response)
except error.HTTPError as e:
LOGGER.debug(e)
def search_pattern(self, file_content, pattern):
"""find strings for CONTAINS_PATTERNS with product_name in them"""
file_content_list = file_content.split("\n")
matches = [
i.strip() for i in file_content_list
if re.search(pattern, i, re.IGNORECASE)
]
LOGGER.debug(f"found matches = {matches}"
) # TODO: regex highlight in these matched strings?
return matches
def find_vendor_product(self):
"""find vendor-product pairs from database"""
LOGGER.debug(
f"checking for product_name='{self.product_name}' and version_name='{self.version_number}' in the database"
)
CVEDB.db_open(self)
cursor = self.connection.cursor()
# finding out all distinct (vendor, product) pairs with the help of product_name
query = """
SELECT distinct vendor, product FROM cve_range
WHERE product=(:product);
"""
cursor.execute(query, {"product": self.product_name})
data = cursor.fetchall()
# checking if (vendor, product) was found in the database
if data:
# warning the user to select the vendor-product pairs manually if multiple pairs are found
if len(data) != 1:
LOGGER.warning(
textwrap.dedent(f"""
===============================================================
Multiple ("vendor", "product") pairs found for "{self.product_name}"
Please manually select the appropriate pair.
===============================================================
"""))
return data # [('vendor', 'product')]
else:
if self.product_name:
# removing numeric characters from the product_name
if any(char.isdigit() for char in self.product_name):
LOGGER.debug(
f"removing digits from product_name={self.product_name}"
)
self.product_name = "".join(
filter(lambda x: not x.isdigit(), self.product_name))
return self.find_vendor_product()
else:
# raise error and ask for product_name
LOGGER.warning(
textwrap.dedent(f"""
=================================================================
No match was found for "{self.product_name}" in database.
Please check your file or try specifying the "product_name" also.
=================================================================
"""))
return []
CVEDB.db_close(self)
def parse_spdx_json(self, sbom_file: str) -> List[List[str]]:
"""parses SPDX JSON BOM file extracting package name and version"""
data = json.load(open(sbom_file))
modules: List[List[str]] = []
for d in data["packages"]:
package = d["name"]
try:
version = d["versionInfo"]
modules.append([package, version])
except KeyError as e:
LOGGER.debug(e, exc_info=True)
return modules
def extract_and_parse_file(self, filename):
"""extracts and parses the file for common patterns, version strings and common filename patterns"""
with self.extractor as ectx:
if ectx.can_extract(filename):
binary_string_list = []
for filepath in self.walker([ectx.extract(filename)]):
clean_path = self.version_scanner.clean_file_path(filepath)
LOGGER.debug(f"checking whether {clean_path} is binary")
# see if the file is ELF binary file and parse for strings
is_exec = self.version_scanner.is_executable(filepath)[0]
if is_exec:
LOGGER.debug(
f"{clean_path} <--- this is an ELF binary")
file_content = self.version_scanner.parse_strings(
filepath)
matches = self.search_pattern(file_content,
self.product_name)
# searching for version strings in the found matches
version_string = self.search_version_string(matches)
self.version_pattern += version_string
# if version string is found in file, append it to filename_pattern
if version_string:
if sys.platform == "win32":
self.filename_pattern.append(
filepath.split("\\")[-1])
else:
self.filename_pattern.append(
filepath.split("/")[-1])
LOGGER.info(
f"matches for {self.product_name} found in {clean_path}"
)
binary_string_list += matches
for i in matches:
if ("/" not in i and "!" not in i
) and len(i) > self.string_length:
self.contains_patterns.append(i)
LOGGER.debug(f"{self.filename_pattern}")
# to resolve case when there are no strings common with product_name in them
if self.contains_patterns:
return self.contains_patterns
return binary_string_list
def scan_file(self) -> Dict[ProductInfo, TriageData]:
LOGGER.info(f"Processing SBOM {self.filename} of type {self.type.upper()}")
try:
if self.type == "spdx":
spdx = SPDXParser()
modules = spdx.parse(self.filename)
elif self.type == "cyclonedx":
cyclone = CycloneParser()
modules = cyclone.parse(self.filename)
elif self.type == "swid":
swid = SWIDParser()
modules = swid.parse(self.filename)
else:
modules = []
except (KeyError, FileNotFoundError, ET.ParseError) as e:
LOGGER.debug(e, exc_info=True)
modules = []
LOGGER.debug(
f"The number of modules identified in SBOM - {len(modules)}\n{modules}"
)
# Now process list of modules to create [vendor, product, version] tuples
parsed_data: List[ProductInfo] = []
for m in modules:
product, version = m[0], m[1]
if version != "":
# Now add vendor to create product record....
# print (f"Find vendor for {product} {version}")
vendor = self.get_vendor(product)
if vendor is not None:
parsed_data.append(ProductInfo(vendor, product, version))
# print(vendor,product,version)
for row in parsed_data:
self.sbom_data[row]["default"] = {
"remarks": Remarks.NewFound,
"comments": "",
"severity": "",
}
self.sbom_data[row]["paths"] = set(map(lambda x: x.strip(), "".split(",")))
LOGGER.debug(f"SBOM Data {self.sbom_data}")
return self.sbom_data
if not self.years():
raise EmptyCache(self.cachedir)
self.LOGGER.debug(f"Years present: {self.years()}")
return self
def __exit__(self, _exc_type, _exc_value, _traceback):
pass
def clear_cached_data(self):
if os.path.exists(self.cachedir):
self.LOGGER.warning(f"Deleting cachedir {self.cachedir}")
shutil.rmtree(self.cachedir)
def refresh():
with CVEDB():
pass
if __name__ == "__main__":
LOGGER.debug("Experimenting...")
cvedb = CVEDB(os.path.join(os.path.expanduser("~"), ".cache", "cvedb"))
# cvedb.refresh()
# print(cvedb.years())
# connection = cvedb.init_database()
# cvedb.populate_db(connection)
# cvedb.supplement_curl()
LOGGER.setLevel(logging.INFO)
LOGGER.info("Getting cves for curl 7.34.0")
LOGGER.info(cvedb.get_cves("haxx", "curl", "7.34.0"))
def main(argv=None):
"""Scan a binary file for certain open source libraries that may have CVEs"""
argv = argv or sys.argv
# Reset logger level to info
LOGGER.setLevel(logging.INFO)
parser = argparse.ArgumentParser(
prog="cve-bin-tool",
description=textwrap.dedent("""
The CVE Binary Tool scans for a number of common, vulnerable open source
components (openssl, libpng, libxml2, expat and a few others) to let you know
if a given directory or binary file includes common libraries with known
vulnerabilities.
"""),
epilog=textwrap.fill(
f'Available checkers: {", ".join(VersionScanner.available_checkers())}'
) + "\n\nPlease disclose issues responsibly!",
formatter_class=argparse.RawDescriptionHelpFormatter,
)
input_group = parser.add_argument_group("Input")
input_group.add_argument("directory",
help="directory to scan",
nargs="?",
default=None)
input_group.add_argument(
"-e",
"--exclude",
action=StringToListAction,
help="Comma separated Exclude directory path",
default=None,
)
input_group.add_argument(
"-i",
"--input-file",
action="store",
default="",
help="provide input filename",
)
input_group.add_argument("-C",
"--config",
action="store",
default="",
help="provide config file")
output_group = parser.add_argument_group("Output")
output_group.add_argument("-q",
"--quiet",
action="store_true",
help="suppress output")
output_group.add_argument(
"-l",
"--log",
help="log level (default: info)",
dest="log_level",
action="store",
choices=["debug", "info", "warning", "error", "critical"],
)
output_group.add_argument(
"-o",
"--output-file",
action="store",
help="provide output filename (default: output to stdout)",
)
output_group.add_argument(
"--html-theme",
action="store",
help="provide custom theme directory for HTML Report",
)
output_group.add_argument(
"-f",
"--format",
action="store",
choices=["csv", "json", "console", "html", "pdf"],
help="update output format (default: console)",
)
output_group.add_argument(
"-c",
"--cvss",
action="store",
help=
"minimum CVSS score (as integer in range 0 to 10) to report (default: 0)",
)
output_group.add_argument(
"-S",
"--severity",
action="store",
choices=["low", "medium", "high", "critical"],
help="minimum CVE severity to report (default: low)",
)
parser.add_argument("-V", "--version", action="version", version=VERSION)
parser.add_argument(
"-u",
"--update",
action="store",
choices=["now", "daily", "never", "latest"],
help="update schedule for NVD database (default: daily)",
)
parser.add_argument(
"-x",
"--extract",
action="store_true",
help="autoextract compressed files",
)
parser.add_argument(
"--disable-version-check",
action="store_true",
help="skips checking for a new version",
)
checker_group = parser.add_argument_group("Checkers")
checker_group.add_argument(
"-s",
"--skips",
dest="skips",
action=StringToListAction,
type=str,
help="comma-separated list of checkers to disable",
)
checker_group.add_argument(
"-r",
"--runs",
dest="runs",
action=StringToListAction,
type=str,
help="comma-separated list of checkers to enable",
)
defaults = {
"directory": "",
"exclude": [],
"input_file": "",
"log_level": "info",
"format": "console",
"cvss": 0,
"severity": "low",
"update": "daily",
"extract": True,
"disable_version_check": False,
"skips": "",
"runs": "",
"quiet": False,
"output_file": "",
"html_theme": "",
}
with ErrorHandler(mode=ErrorMode.NoTrace):
raw_args = parser.parse_args(argv[1:])
args = {key: value for key, value in vars(raw_args).items() if value}
configs = {}
if args.get("config"):
conf = ConfigParser(args["config"])
configs = conf.parse_config()
args = ChainMap(args, configs, defaults)
# logging and error related settings
if args["log_level"]:
LOGGER.setLevel(args["log_level"].upper())
if args["quiet"]:
LOGGER.setLevel(logging.CRITICAL)
if 0 < LOGGER.level <= 10:
error_mode = ErrorMode.FullTrace
elif LOGGER.level >= 50:
error_mode = ErrorMode.NoTrace
else:
error_mode = ErrorMode.TruncTrace
if platform.system() != "Linux":
warning_nolinux = """
**********************************************
Warning: this utility was developed for Linux.
You may need to install additional utilities
to use it on other operating systems.
**********************************************
"""
LOGGER.warning(warning_nolinux)
# Database update related settings
# Connect to the database
cvedb_orig = CVEDB(version_check=not args["disable_version_check"],
error_mode=error_mode)
# if OLD_CACHE_DIR (from cvedb.py) exists, print warning
if os.path.exists(OLD_CACHE_DIR):
LOGGER.warning(
f"Obsolete cache dir {OLD_CACHE_DIR} is no longer needed and can be removed."
)
# Clear data if -u now is set
if args["update"] == "now":
cvedb_orig.clear_cached_data()
if args["update"] == "latest":
cvedb_orig.refresh_cache_and_update_db()
# update db if needed
if args["update"] != "never":
cvedb_orig.get_cvelist_if_stale()
else:
LOGGER.warning("Not verifying CVE DB cache")
if not cvedb_orig.nvd_years():
with ErrorHandler(mode=error_mode, logger=LOGGER):
raise EmptyCache(cvedb_orig.cachedir)
# CVE Database validation
if not cvedb_orig.check_cve_entries():
with ErrorHandler(mode=error_mode, logger=LOGGER):
raise CVEDataMissing("No data in CVE Database")
# Input validation
if not args["directory"] and not args["input_file"]:
parser.print_usage()
with ErrorHandler(logger=LOGGER, mode=ErrorMode.NoTrace):
raise InsufficientArgs(
"Please specify a directory to scan or an input file required")
if args["directory"] and not os.path.exists(args["directory"]):
parser.print_usage()
with ErrorHandler(logger=LOGGER, mode=ErrorMode.NoTrace):
raise FileNotFoundError("Directory/File doesn't exist")
# Checkers related settings
skips = args["skips"]
if args["runs"]:
runs = args["runs"]
skips = list(
map(
lambda checker: checker.name,
filter(
lambda checker: checker.name not in runs,
pkg_resources.iter_entry_points("cve_bin_tool.checker"),
),
))
# CSVScanner related settings
score = 0
if args["severity"]:
# Set minimum CVSS score based on severity
cvss_score = {"low": 0, "medium": 4, "high": 7, "critical": 9}
score = cvss_score[args["severity"]]
if int(args["cvss"]) > 0:
score = int(args["cvss"])
with CVEScanner(score=score) as cve_scanner:
triage_data: TriageData
total_files: int = 0
parsed_data: Dict[ProductInfo, TriageData] = {}
if args["input_file"]:
input_engine = InputEngine(args["input_file"],
logger=LOGGER,
error_mode=error_mode)
parsed_data = input_engine.parse_input()
if not args["directory"]:
for product_info, triage_data in parsed_data.items():
LOGGER.warning(f"{product_info}, {triage_data}")
cve_scanner.get_cves(product_info, triage_data)
if args["directory"]:
version_scanner = VersionScanner(
should_extract=args["extract"],
exclude_folders=args["exclude"],
error_mode=error_mode,
)
version_scanner.remove_skiplist(skips)
version_scanner.print_checkers()
for scan_info in version_scanner.recursive_scan(args["directory"]):
if scan_info:
product_info, path = scan_info
LOGGER.debug(f"{product_info}: {path}")
triage_data = parsed_data.get(product_info,
{"default": {}})
# Ignore paths from triage_data if we are scanning directory
triage_data["paths"] = {path}
cve_scanner.get_cves(product_info, triage_data)
total_files = version_scanner.total_scanned_files
LOGGER.info("")
LOGGER.info("Overall CVE summary: ")
if args["input_file"]:
LOGGER.info(
f"There are {cve_scanner.products_with_cve} products with known CVEs detected"
)
else:
LOGGER.info(
f"There are {cve_scanner.products_with_cve} files with known CVEs detected"
)
if cve_scanner.products_with_cve > 0 or (args["format"] == "html"
or args["format"] == "pdf"):
affected_string = ", ".join(
map(
lambda product_version: "".join(str(product_version)),
cve_scanner.affected(),
))
LOGGER.info(f"Known CVEs in {affected_string}:")
# Creates a Object for OutputEngine
output = OutputEngine(
all_cve_data=cve_scanner.all_cve_data,
scanned_dir=args["directory"],
filename=args["output_file"],
themes_dir=args["html_theme"],
products_with_cve=cve_scanner.products_with_cve,
products_without_cve=cve_scanner.products_without_cve,
total_files=total_files,
)
if not args["quiet"]:
output.output_file(args["format"])
# Use the number of products with known cves as error code
# as requested by folk planning to automate use of this script.
# If no cves found, then the program exits cleanly.
return cve_scanner.products_with_cve
def main(argv=None):
argv = argv or sys.argv
parser = argparse.ArgumentParser(
prog="helper-script",
description=textwrap.dedent("""
Helps contributors who want to write a new cve-bin-tool checker find common filenames,
version strings, and other necessary data for building a binary checker
"""),
)
# scan directory args
parser.add_argument(
"filenames",
help="files to scan",
nargs="+",
default=[],
)
# product-name args
parser.add_argument(
"-p",
"--product",
help="provide product-name that would be searched",
dest="product_name",
action="store",
default=None,
)
# version-name args
parser.add_argument(
"-v",
"--version",
help="provide version that would be searched",
dest="version_number",
action="store",
default=None,
)
# log level args
parser.add_argument(
"-l",
"--log",
help="log level (default: warning)",
dest="log_level",
action="store",
choices=["debug", "info", "warning", "error", "critical"],
default="warning",
)
# contains-patterns string length args
parser.add_argument(
"--string-length",
help=
"changes the output string-length for CONTAINS_PATTERNS (default: %(default)s)",
type=int,
action="store",
default=40,
)
with ErrorHandler(mode=ErrorMode.NoTrace):
raw_args = parser.parse_args(argv[1:])
args = {key: value for key, value in vars(raw_args).items() if value}
defaults = {key: parser.get_default(key) for key in vars(raw_args)}
args = ChainMap(args, defaults)
LOGGER.setLevel(args["log_level"].upper())
LOGGER.debug(f"Given filenames: {args['filenames']}")
LOGGER.info(f"Scanning only the first filename: '{args['filenames'][0]}'")
hs = HelperScript(
args["filenames"][0],
product_name=args["product_name"],
version_number=args["version_number"],
string_length=args["string_length"],
)
# Parsing, Extracting and Searching for version-strings
hs.extract_and_parse_file(args["filenames"][0])
# output on console
hs.output()
