Class: Arachni::Element::LinkTemplate

Inherits:
Base show all
Includes:
Capabilities::Analyzable, Capabilities::WithDOM, Capabilities::WithNode
Defined in:
lib/arachni/element/link_template.rb,
lib/arachni/element/link_template/dom.rb

Overview

Represents an auditable link element whose inputs have been extracted based on an externally provided template.

Author:

Defined Under Namespace

Classes: DOM

Constant Summary collapse

INVALID_INPUT_DATA = 
[
    # Protocol URLs require a // which we can't preserve.
    '://'
]

Constants included from Capabilities::Analyzable::Differential

Capabilities::Analyzable::Differential::DIFFERENTIAL_OPTIONS

Constants included from Capabilities::Analyzable::Taint

Capabilities::Analyzable::Taint::TAINT_OPTIONS

Constants included from Capabilities::Auditable

Capabilities::Auditable::OPTIONS

Constants included from Capabilities::Mutable

Capabilities::Mutable::MUTATION_OPTIONS

Instance Attribute Summary collapse

Attributes included from Capabilities::Auditable

#audit_options

Attributes included from Capabilities::WithAuditor

#auditor

Attributes included from Capabilities::Mutable

#affected_input_name, #format, #seed

Attributes included from Capabilities::Inputtable

#default_inputs, #inputs

Attributes included from Capabilities::WithNode

#html

Attributes inherited from Base

#initialization_options, #page

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Capabilities::Analyzable

has_timeout_candidates?, reset, timeout_audit_run

Methods included from Capabilities::Analyzable::Differential

#differential_analysis, reset

Methods included from Capabilities::Analyzable::Timeout

add_phase_2_candidate, candidates_include?, deduplicate, deduplicate?, do_not_deduplicate, #ensure_responsiveness, has_candidates?, payload_delay_from_options, reset, run, #timeout_analysis, timeout_from_options, #timeout_id, #timing_attack_probe, #timing_attack_verify

Methods included from Capabilities::Analyzable::Taint

#taint_analysis

Methods included from Capabilities::Auditable

#audit, #audit_id, #audit_status_message, #audit_status_message_action, #audit_verbose_message, #coverage_hash, #dup, #matches_skip_like_blocks?, #reset, reset, #skip?, skip_like

Methods included from Capabilities::WithAuditor

#dup, #marshal_dump, #orphan?, #prepare_for_report, #remove_auditor

Methods included from Capabilities::Mutable

#affected_input_value, #affected_input_value=, #dup, #each_mutation, #immutables, #mutation?, #mutations, #reset, #switch_method, #to_h

Methods included from Capabilities::Submittable

#action, #action=, #dup, #http, #method, #method=, #platforms, #submit, #to_h

Methods included from Capabilities::Inputtable

#[], #[]=, #changes, #dup, #has_inputs?, #inputtable_id, #reset, #to_h, #try_input, #update, #valid_input_name_data?, #valid_input_value?, #valid_input_value_data?

Methods included from Utilities

#available_port, #caller_name, #caller_path, #cookie_decode, #cookie_encode, #cookies_from_document, #cookies_from_file, #cookies_from_response, #exception_jail, #exclude_path?, #follow_protocol?, #form_decode, #form_encode, #forms_from_document, #forms_from_response, #generate_token, #get_path, #hms_to_seconds, #html_decode, #html_encode, #include_path?, #links_from_document, #links_from_response, #normalize_url, #page_from_response, #page_from_url, #parse_set_cookie, #path_in_domain?, #path_too_deep?, #port_available?, #rand_port, #random_seed, #redundant_path?, #remove_constants, #request_parse_body, #seconds_to_hms, #skip_page?, #skip_path?, #skip_resource?, #skip_response?, #to_absolute, #uri_decode, #uri_encode, #uri_parse, #uri_parse_query, #uri_parser, #uri_rewrite

Methods included from Capabilities::WithDOM

#dup

Methods included from Capabilities::WithNode

#dup, #node, #to_h

Methods inherited from Base

#==, #action, #dup, #hash, #marshal_dump, #marshal_load, #persistent_hash, #prepare_for_report, #reset, #to_h, #to_hash, #type, #url, #url=

Methods included from Capabilities::WithScope

#scope

Constructor Details

#initialize(options) ⇒ LinkTemplate

Returns a new instance of LinkTemplate.

Parameters:

Options Hash (options):

  • :url (String) —

    URL of the page which includes the link.

  • :action (String) —

    Link URL – defaults to ‘:url`.

  • :inputs (Hash) —

    Query parameters as ‘name => value` pairs. If none have been provided they will automatically be extracted from Capabilities::Submittable#action.



43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
 
# File 'lib/arachni/element/link_template.rb', line 43

def initialize( options )
    super( options )

    @template = options[:template]

    if options[:inputs]
        self.inputs = options[:inputs]
    else
        if @template
            _, inputs = self.class.extract_inputs( self.action, [@template] )
            self.inputs = inputs if inputs
        else
            @template, inputs = self.class.extract_inputs( self.action )

            if @template
                self.inputs = inputs
            end
        end
    end

    self.inputs ||= {}
    @default_inputs = self.inputs.dup.freeze
end

Instance Attribute Details

#template ⇒ Regexp (readonly)

Returns Regular expressions with named captures, serving as templates used to identify and manipulate inputs in Capabilities::Submittable#action.

Returns:



33
34
35
 
# File 'lib/arachni/element/link_template.rb', line 33

def template
  @template
end

Class Method Details

.decode(*args) ⇒ Object 



238
239
240
 
# File 'lib/arachni/element/link_template.rb', line 238

def decode( *args )
    URI.decode( *args )
end

.encode(string) ⇒ Object 



234
235
236
 
# File 'lib/arachni/element/link_template.rb', line 234

def encode( string )
    URI.encode( URI.encode( URI.encode( string, ';' ) ), '/' )
end

.extract_inputs(url, templates = Arachni::Options.audit.link_templates) ⇒ Array[Regexp, Hash]

Extracts inputs from a URL based on the given templates.

Parameters:

  • url (String)
  • templates (Array<Regexp>) (defaults to: Arachni::Options.audit.link_templates)

Returns:

  • (Array[Regexp, Hash]) —

    Matched template and inputs.



219
220
221
222
223
224
225
226
227
228
229
230
231
232
 
# File 'lib/arachni/element/link_template.rb', line 219

def extract_inputs( url, templates = Arachni::Options.audit.link_templates )
    return [] if !url || templates.empty?

    templates.each do |template|
        if (match = url.scan_in_groups( template )).any?
            return [
                template,
                match.inject({}){ |h, (k, v)| h[k] = decode(v); h }
            ]
        end
    end

    []
end

.from_document(url, document, templates = Arachni::Options.audit.link_templates) ⇒ Array<LinkTemplate>

Extracts link with inputs based on the provided templates from a document.

Parameters:

  • url (String) —

    URL of the document – used for path normalization purposes.

  • document (String, Nokogiri::HTML::Document)
  • templates (Array<Regexp>) (defaults to: Arachni::Options.audit.link_templates)

Returns:



182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
 
# File 'lib/arachni/element/link_template.rb', line 182

def from_document( url, document, templates = Arachni::Options.audit.link_templates )
    return [] if templates.empty?

    document = Nokogiri::HTML( document.to_s ) if !document.is_a?( Nokogiri::HTML::Document )
    base_url = begin
        document.search( '//base[@href]' )[0]['href']
    rescue
        url
    end

    document.search( '//a' ).map do |link|
        next if !(href = to_absolute( link['href'], base_url ))

        template, inputs = extract_inputs( href, templates )
        next if !template && !DOM.data_from_node( link )

        if (parsed_url = Arachni::URI( href ))
            next if parsed_url.scope.out?
        end

        new(
            url:      url.freeze,
            action:   href.freeze,
            inputs:   inputs || {},
            template: template,
            html:     link.to_html.freeze
        )
    end.compact
end

.from_response(response, templates = Arachni::Options.audit.link_templates) ⇒ Array<Link>

Extracts links from an HTTP response.

Parameters:

Returns:



154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
 
# File 'lib/arachni/element/link_template.rb', line 154

def from_response( response, templates = Arachni::Options.audit.link_templates )
    url = response.url

    links = from_document( url, response.body, templates )

    template, inputs = extract_inputs( url, templates )
    if template
        links << new(
            url:      url.freeze,
            action:   url.freeze,
            inputs:   inputs,
            template: template
        )
    end

    links
end

.from_rpc_data(data) ⇒ Object 



143
144
145
146
 
# File 'lib/arachni/element/link_template.rb', line 143

def from_rpc_data( data )
    data['template'] = Regexp.new( data['template'] ) if data['template']
    super data
end

.type ⇒ Object 



242
243
244
 
# File 'lib/arachni/element/link_template.rb', line 242

def type
    :link_template
end

Instance Method Details

#coverage_id ⇒ Object 



123
124
125
 
# File 'lib/arachni/element/link_template.rb', line 123

def coverage_id
    dom_data ? "#{super}:#{dom_data[:inputs].keys.sort}" : super
end

#decode(*args) ⇒ Object 



119
120
121
 
# File 'lib/arachni/element/link_template.rb', line 119

def decode( *args )
    self.class.decode( *args )
end

#dom ⇒ DOM

Returns:



68
69
70
71
72
73
 
# File 'lib/arachni/element/link_template.rb', line 68

def dom
    return @dom if @dom
    return if !dom_data

    super
end

#encode(string) ⇒ Object 



115
116
117
 
# File 'lib/arachni/element/link_template.rb', line 115

def encode( string )
    self.class.encode( string )
end

#id ⇒ Object 



127
128
129
 
# File 'lib/arachni/element/link_template.rb', line 127

def id
    dom_data ? "#{super}:#{dom_data[:inputs].sort_by { |k,_| k }}" : super
end

#simple ⇒ Hash

Returns Simple representation of self in the form of ‘{ Capabilities::Submittable#action => Capabilities::Inputtable#inputs }`.

Returns:



100
101
102
 
# File 'lib/arachni/element/link_template.rb', line 100

def simple
    { self.action => self.inputs }
end

#to_rpc_data ⇒ Object 



131
132
133
134
135
136
137
138
139
 
# File 'lib/arachni/element/link_template.rb', line 131

def to_rpc_data
    data = super
    return data if !@template

    data.merge!( 'template' => @template.source )
    data['initialization_options'][:template] = data['template']
    data.delete 'dom_data'
    data
end

#to_s ⇒ String

Returns URL updated with the configured Capabilities::Inputtable#inputs.

Returns:



106
107
108
109
110
111
112
113
 
# File 'lib/arachni/element/link_template.rb', line 106

def to_s
    return self.action if self.inputs.empty?

    self.action.sub_in_groups(
        @template,
        inputs.inject({}) { |h, (k, v)| h[k] = encode(v); h }
    )
end

#valid_input_data?(data) ⇒ Bool

Returns ‘true` if the `data` don’t contain strings specified in #INVALID_INPUT_DATA, ‘false` otherwise.

Parameters:

  • data (String) —

    Input data.

Returns:

  • (Bool) —

    ‘true` if the `data` don’t contain strings specified in #INVALID_INPUT_DATA, ‘false` otherwise.

See Also:



94
95
96
 
# File 'lib/arachni/element/link_template.rb', line 94

def valid_input_data?( data )
    !INVALID_INPUT_DATA.find { |c| data.include? c }
end

#valid_input_name?(name) ⇒ Bool

Returns ‘true` if the `name` can be found as a named capture in #template, `false` otherwise.

Parameters:

  • name (String) —

    Input name.

Returns:

  • (Bool) —

    ‘true` if the `name` can be found as a named capture in #template, `false` otherwise.



81
82
83
84
 
# File 'lib/arachni/element/link_template.rb', line 81

def valid_input_name?( name )
    return if !@template
    @template.names.include? name
end