metasploit-framework/lib/rex/parser/appscan_nokogiri.rb

367 lines
11 KiB
Ruby

require "rex/parser/nokogiri_doc_mixin"
module Rex
module Parser
# If Nokogiri is available, define AppScan document class.
load_nokogiri && class AppscanDocument < Nokogiri::XML::SAX::Document
include NokogiriDocMixin
# The resolver prefers your local /etc/hosts (or windows equiv), but will
# fall back to regular DNS. It retains a cache for the import to avoid
# spamming your network with DNS requests.
attr_reader :resolv_cache
# If name resolution of the host fails out completely, you will not be
# able to import that Scan task. Other scan tasks in the same report
# should be unaffected.
attr_reader :parse_warning
def start_document
@parse_warnings = []
@resolv_cache = {}
end
def start_element(name=nil,attrs=[])
attrs = normalize_attrs(attrs)
block = @block
@state[:current_tag][name] = true
case name
when "Issue" # Start of the stuff we want
collect_issue(attrs)
when "Entity" # Start of the stuff we want
collect_entity(attrs)
when "Severity", "Url", "OriginalHttpTraffic"
@state[:has_text] = true
end
end
def end_element(name=nil)
block = @block
case name
when "Issue" # Wrap it up
record_issue
# Reset the state once we close an issue
@state = @state.select do
|k| [:current_tag, :web_sites].include? k
end
when "Url" # Populates @state[:web_site]
@state[:has_text] = false
record_url
@text = nil
report_web_site(&block)
handle_parse_warnings(&block)
when "Severity"
@state[:has_text] = false
record_risk
@text = nil
when "OriginalHttpTraffic" # Request and response
@state[:has_text] = false
record_request_and_response
report_service_info
page_info = report_web_page(&block)
if page_info
form_info = report_web_form(page_info,&block)
if form_info
report_web_vuln(form_info,&block)
end
end
@text = nil
end
@state[:current_tag].delete name
end
def report_web_vuln(form_info,&block)
return unless(in_issue && has_text)
return unless form_info.kind_of? Hash
return unless @state[:issue]
return unless @state[:issue]["Noise"]
return unless @state[:issue]["Noise"].to_s.downcase == "false"
return unless @state[:issue][:vuln_param]
web_vuln_info = {}
web_vuln_info[:web_site] = form_info[:web_site]
web_vuln_info[:path] = form_info[:path]
web_vuln_info[:query] = form_info[:query]
web_vuln_info[:method] = form_info[:method]
web_vuln_info[:params] = form_info[:params]
web_vuln_info[:pname] = @state[:issue][:vuln_param]
web_vuln_info[:proof] = "" # TODO: pick this up from <Difference> maybe?
web_vuln_info[:risk] = @state[:issue][:risk]
web_vuln_info[:name] = @state[:issue]["IssueTypeID"]
web_vuln_info[:category] = "imported"
web_vuln_info[:confidence] = 100 # Seems pretty binary, noise or not
db.emit(:web_vuln, web_vuln_info[:name], &block) if block
web_vuln = db_report(:web_vuln, web_vuln_info)
end
def collect_entity(attrs)
return unless in_issue
return unless @state[:issue].kind_of? Hash
ent_hash = attr_hash(attrs)
return unless ent_hash
return unless ent_hash["Type"].to_s.downcase == "parameter"
@state[:issue][:vuln_param] = ent_hash["Name"]
end
def report_web_form(page_info,&block)
return unless(in_issue && has_text)
return unless page_info.kind_of? Hash
return unless @state[:request_body]
return if @state[:request_body].strip.empty?
web_form_info = {}
web_form_info[:web_site] = page_info[:web_site]
web_form_info[:path] = page_info[:path]
web_form_info[:query] = page_info[:query]
web_form_info[:method] = @state[:request_headers].cmd_string.split(/\s+/)[0]
parsed_params = parse_params(@state[:request_body])
return unless parsed_params
return if parsed_params.empty?
web_form_info[:params] = parsed_params
web_form = db_report(:web_form, web_form_info)
@state[:web_forms] ||= []
unless @state[:web_forms].include? web_form
db.emit(:web_form, @state[:uri].to_s, &block) if block
@state[:web_forms] << web_form
end
web_form_info
end
def parse_params(request_body)
return unless request_body
pairs = request_body.split(/&/)
params = []
pairs.each do |pair|
param,value = pair.split("=",2)
params << [param,""] # Can't tell what's default
end
params
end
def report_web_page(&block)
return unless(in_issue && has_text)
return unless @state[:web_site]
return unless @state[:response_headers]
return unless @state[:uri]
web_page_info = {}
web_page_info[:web_site] = @state[:web_site]
web_page_info[:path] = @state[:uri].path
web_page_info[:body] = @state[:response_body].to_s
web_page_info[:query] = @state[:uri].query
code = @state[:response_headers].cmd_string.split(/\s+/)[1]
return unless code
web_page_info[:code] = code
parsed_headers = {}
@state[:response_headers].each do |k,v|
parsed_headers[k.to_s.downcase] ||= []
parsed_headers[k.to_s.downcase] << v
end
return if parsed_headers.empty?
web_page_info[:headers] = parsed_headers
web_page = db_report(:web_page, web_page_info)
@state[:web_pages] ||= []
unless @state[:web_pages].include? web_page
db.emit(:web_page, @state[:uri].to_s, &block) if block
@state[:web_pages] << web_page
end
web_page_info
end
def report_service_info
return unless(in_issue && has_text)
return unless @state[:web_site]
return unless @state[:response_headers]
banner = @state[:response_headers]["server"]
return unless banner
service = @state[:web_site].service
return unless service.info.to_s.empty?
service_info = {
:host => service.host,
:port => service.port,
:proto => service.proto,
:info => banner
}
db_report(:service, service_info)
end
def record_request_and_response
return unless(in_issue && has_text)
return unless @state[:web_site]
really_original_traffic = unindent_and_crlf(@text)
split_traffic = really_original_traffic.split(/\r\n\r\n/)
request_headers_text = split_traffic.first
content_length = 0
if request_headers_text =~ /\ncontent-length:\s+([0-9]+)/mni
content_length = $1.to_i
end
if(content_length > 0) and (split_traffic[1].to_s.size >= content_length)
request_body_text = split_traffic[1].to_s[0,content_length]
else
request_body_text = nil
end
response_headers_text = split_traffic[1].to_s[content_length,split_traffic[1].to_s.size].lstrip
request = request_headers_text
return unless(request && response_headers_text)
response_body_text = split_traffic[2]
req_header = Rex::Proto::Http::Packet::Header.new
res_header = Rex::Proto::Http::Packet::Header.new
req_header.from_s request_headers_text.dup
res_header.from_s response_headers_text.dup
@state[:request_headers] = req_header
@state[:request_body] = request_body_text
@state[:response_headers] = res_header
@state[:response_body] = response_body_text
end
# Appscan tab-indents which makes parsing a little difficult. They
# also don't record CRLFs, just LFs.
def unindent_and_crlf(text)
second_line = text.split(/\r*\n/)[1]
indent_level = second_line.size - second_line.lstrip.size
unindented_text_lines = []
text.split(/\r*\n/).each do |line|
if line =~ /^\t{#{indent_level}}/
unindented_line = line[indent_level,line.size]
unindented_text_lines << unindented_line
else
unindented_text_lines << line
end
end
unindented_text_lines.join("\r\n")
end
def record_risk
return unless(in_issue && has_text)
@state[:issue] ||= {}
@state[:issue][:risk] = map_severity_to_risk
end
def map_severity_to_risk
case @text.to_s.downcase
when "high" ; 5
when "medium" ; 3
when "low" ; 1
else ; 0
end
end
# TODO
def record_issue
return unless in_issue
return unless @report_data[:issue].kind_of? Hash
return unless @state[:web_site]
return if @state[:issue]["Noise"].to_s.downcase == "true"
end
def collect_issue(attrs)
return unless in_issue
@state[:issue] = {}
@state[:issue].merge! attr_hash(attrs)
end
def report_web_site(&block)
return unless @state[:uri]
uri = @state[:uri]
hostname = uri.host # Assume the first one is the real hostname
address = resolve_issue_url_address(uri)
return unless address
unless @resolv_cache.values.include? address
db.emit(:address, address, &block) if block
end
port = resolve_port(uri)
return unless port
scheme = uri.scheme
return unless scheme
web_site_info = {:workspace => @args[:wspace]}
web_site_info[:vhost] = hostname
service_obj = check_for_existing_service(address,port)
if service_obj
web_site_info[:service] = service_obj
else
web_site_info[:host] = address
web_site_info[:port] = port
web_site_info[:ssl] = scheme == "https"
end
web_site_obj = db_report(:web_site, web_site_info)
@state[:web_sites] ||= []
unless @state[:web_sites].include? web_site_obj
url = "#{uri.scheme}://#{uri.host}:#{uri.port}"
db.emit(:web_site, url, &block) if block
db.report_import_note(@args[:wspace], web_site_obj.service.host)
@state[:web_sites] << web_site_obj
end
@state[:service] = service_obj || web_site_obj.service
@state[:host] = (service_obj || web_site_obj.service).host
@state[:web_site] = web_site_obj
end
def check_for_existing_service(address,port)
db.get_service(@args[:wspace],address,"tcp",port)
end
def resolve_port(uri)
@state[:port] = uri.port
unless @state[:port]
@parse_warnings << "Could not determine a port for '#{@state[:scan_name]}'"
end
return @state[:port]
end
def resolve_address(host)
return @resolv_cache[host] if @resolv_cache[host]
address = Rex::Socket.resolv_to_dotted(host) rescue nil
@resolv_cache[host] = address
if address
block = @block
db.emit(:address, address, &block) if block
end
return address
end
# Alias this
def resolve_issue_url_address(uri)
if uri.host
address = resolve_address(uri.host)
unless address
@parse_warnings << "Could not resolve address for '#{uri.host}', skipping."
end
else
@parse_warnings << "Could not determine a host for this import."
end
address
end
def handle_parse_warnings(&block)
return if @parse_warnings.empty?
@parse_warnings.each do |pwarn|
db.emit(:warning, pwarn, &block) if block
end
end
def record_url
return unless in_issue
return unless has_text
uri = URI.parse(@text) rescue nil
return unless uri
@state[:uri] = uri
end
def in_issue
return false unless in_tag("Issue")
return false unless in_tag("Issues")
return false unless in_tag("XmlReport")
return true
end
def has_text
return false unless @text
return false if @text.strip.empty?
@text = @text.strip
end
end
end
end