Add CookieJar support to http_client by agalway-r7 · Pull Request #14831 · rapid7/metasploit-framework

agalway-r7 · 2021-03-01T18:14:21Z

Adds the http-cookie gem, and implements HTTP::CookieJar in http_client. http_client will now automatically collect Set-Cookie responses from the server and include them in future requests. Supersedes #14139.

Calls to send_request_cgi can now include an instance of HTTP::CookieJar in the opts Hash. If included, the passed CookieJar will be used in the request instead of the cookies automatically collected from server responses by http_client.

Calls can also be made to http_client to alter the contents of the module HTTP::CookieJar instance by merging in another CookieJar. This will affect the cookies included in all future requests.

Verification

List the steps needed to make sure this thing works

Start msfconsole
Set up a gitlab test environment
Test execution of gitlab_file_read_rce on gitlab docker instance
See if the module gets a session correctly

modules/exploits/multi/http/gitlab_file_read_rce.rb

dwelch-r7 · 2021-03-02T11:10:55Z

modules/exploits/multi/http/gitlab_file_read_rce.rb

    end

    def sign_in(username, password)
+      http_client.clear_cookies


this worries me a little, why do we need to clear cookies here?
are all modules expected to do this?
do cookie values persist between modules?

You need to clear cookies at this part of the code since you're starting a new communication with the server, and there shouldn't be any cookies in use from previous requests.

Ideally, http_client should be an object instead of a mixin and you'd just create a new http_client for each conversation that needs it's own set of cookies. However, since we haven't done the work to get http_client working as an object yet, clear_cookies has got to stick around as a stopgap.

That being said, I'll double check now to see if cookies remain in the jar between modules, since that could be a problem.

Maybe this functionality should exist in prepend Msf::Exploit::Remote::AutoCheck, this will continue to bite module developers i'd imagine

Random internal thought: I'm surprised this hasn't caused issues sooner with other modules that have state stored via mixins between the check and run methods 🤔

Agreed re: the automatic cookie clear after check, since check should be independent of exploit.

As for why this hasn't happened before, my guess would be that if mixins caused issues moving between check & exploit then it would have been addressed either when the module was first made or when it was landed.

dwelch-r7 · 2021-03-02T11:25:44Z

modules/exploits/multi/http/gitlab_file_read_rce.rb

+    cj = HTTP::CookieJar.new
+
+    cookie_jar.each do |c|
+      next unless c.name == 'experimentation_subject_id'
+
+      payload_cookie = make_default_cookie('experimentation_subject_id', signed_cookie_value)
+
+      payload_cookie.expires = c.expires
+      payload_cookie.max_age = c.max_age
+      payload_cookie.created_at = c.created_at
+      payload_cookie.accessed_at = c.accessed_at
+
+      cj.add(payload_cookie)
+    end


why is there a need for all this? is there not way to just update the cookies value?
or just create it like was done previously? I don't think all this is necessary

There are 3 use cases for interacting with the http_client cookie_jar:

Automatic - cookie_jar collects all cookies sent from the server and applies them to all requests going forward, where appropriate.

One Off - What you see above, you want cookie(s) used in a request, so you pass said cookies as a CookieJar

Global - Use the http_client merge methods to combine a user created CookieJar with the http_client cookie_jar, and said cookies will be applied to all future requests.

Since we're only using these cookies for one request, we don't want to merge them in the http_client cookie_jar, hence the One-Off.

You could make a http_client method that updates cookie values, but it got complicated when you had to account for multiple cookies with the same name but slightly different domains & paths.

lib/msf/core/exploit/remote/http_client.rb

dwelch-r7 · 2021-03-02T12:32:48Z

lib/msf/core/exploit/remote/http_client.rb

+    common_cj = HTTP::CookieJar.new
+    cookie_jar.each { |c| common_cj.add(c) }


Rather than creating a new jar and looping over the old one I think you can use this https://www.rubydoc.info/gems/http-cookie/1.0.3/HTTP/CookieJar#initialize_copy-instance_method

Although I'm genuinely surprised there is no built in merge functionality

I didn't go with the copy method because "Not all backend store classes support cloning." and I wasn't sure what other backends are used in the code, or what might change in the future. Above code is smellier but IMO more future proof

Also yes, no merge is annoying. I think it's because HTTP::CookieJar is just a wrapper around a @store object, which can be a few types of collections. So HTTP::CookieJar has to be sparse with the methods it implements to ensure everything works with all the different @stores

adfoster-r7 · 2021-03-02T15:28:15Z

Before speaking about the specifics of the implementation, I think it would be great to convert over some more modules that are manually keeping track of cookies to see all of the use cases could be handled as well as verifying a few of the recently landed http modules that currently set keep_cookies. The main reason being that for this one module alone we spotted some gremlins already, I imagine there's more that might be a road blocker to a particular approach

Let me know your thoughts 👍

agalway-r7 · 2021-03-02T16:06:04Z

Before speaking about the specifics of the implementation, I think it would be great to convert over some more modules that are manually keeping track of cookies to see all of the use cases could be handled as well as verifying a few of the recently landed http modules that currently set keep_cookies. The main reason being that for this one module alone we spotted some gremlins already, I imagine there's more that might be a road blocker to a particular approach

Let me know your thoughts 👍

Makes sense to me 👍

wvu · 2021-03-04T17:48:35Z

This is great! Thanks for doing this. It incorporates a new gem, but it's way better than the bespoke approach, which is admittedly kinda shoddy.

agalway-r7 · 2021-03-09T10:21:17Z

This is great! Thanks for doing this. It incorporates a new gem, but it's way better than the bespoke approach, which is admittedly kinda shoddy.

I did try to implement the existing cookie-jar gem we already have added (cookiejar), but it's API is not not not good

adfoster-r7 · 2021-03-15T11:07:54Z

lib/msf/core/exploit/remote/http_client.rb


    # Initialize an empty cookie jar to keep cookies
-    self.cookie_jar = Set.new
+    self.cookie_jar = HTTP::CookieJar.new


As discussed on a chat, let's investigate how a wrapper would work:

Suggested change

self.cookie_jar = HTTP::CookieJar.new

self.cookie_jar = Exploit::Remote::HttpCookieJar.new

This will make it easier to unit tests too 🎉

adfoster-r7 · 2021-03-15T11:17:51Z

modules/exploits/multi/http/gitlab_file_read_rce.rb

-    signed_cookie = sign_payload(secret_key_base, payload)
+    signed_cookie_value = sign_payload(secret_key_base, payload)
+
+    cj = HTTP::CookieJar.new


cookie_jar is already available:

Suggested change

cj = HTTP::CookieJar.new

cj = self.cookie_jar

Which actually makes this line redundant, as we can just use cookie_jar instead of the more cryptic cj 🕵️

So with the API we're aiming for I think this ends up being:

cookie_jar.clear cookie_jar.set('experimentation_subject_id', signed_cookie_value)

Thining ahead and future proofing an API for a module were we don't have the luxury of clearing the entire cookie jar by itself, we could also use:

cookie_jar = Exploit::Remote::HttpCookieJar.new cookie_jar.set('experimentation_subject_id', signed_cookie_value)

adfoster-r7 · 2021-03-15T11:18:28Z

modules/exploits/multi/http/gitlab_file_read_rce.rb

+    cookie_jar.each do |c|
+      next unless c.name == 'experimentation_subject_id'
+
+      payload_cookie = make_default_cookie('experimentation_subject_id', signed_cookie_value)
+
+      payload_cookie.expires = c.expires
+      payload_cookie.max_age = c.max_age
+      payload_cookie.created_at = c.created_at
+      payload_cookie.accessed_at = c.accessed_at
+
+      cj.add(payload_cookie)
+    end


This would be a good chunk of code to extract into the http cookie jar wrapper; The python requests API is nice:
https://stackoverflow.com/questions/17224054/how-to-add-a-cookie-to-the-cookiejar-in-python-requests-library

adfoster-r7 · 2021-03-15T11:37:20Z

lib/msf/core/exploit/remote/http_client.rb

+  # Empties the contents of +cookie_jar+, preventing any cookies
+  # currently stored in the jar from being used in future requests.
+  #
+  # @return [void]
+  def clear_cookies
+    cookie_jar.clear
+  end
+
+  # Merges the contents of the +cj+ param into the +cookie_jar+ class var.
+  # In the case of Cookies with the same name, domain, and path, the +cj+ cookie will be used.
+  #
+  # @param cj [HTTP::CookieJar] CookieJar to be merged into +cookie_jar+.
+  #
+  # @return [HTTP::CookieJar] cookie_jar
+  def merge_cookie_jar!(cj)
+    cj.cookies.each do |c|
+      cookie_jar.add(c)
+    end
+
+    cookie_jar
+  end
+
+  # Merges the contents of  +cj+ & +cookie_jar+ into a new CookieJar.
+  # In the case of Cookies with the same name, domain, and path, the +cj+ cookie will be used.
+  #
+  # @param cj [HTTP::CookieJar] CookieJar to be merged into a new CookieJar with +cookie_jar+.
+  #
+  # @return [HTTP::CookieJar]
+  def merge_cookie_jar(cj)
+    common_cj = HTTP::CookieJar.new
+    cookie_jar.each { |c| common_cj.add(c) }
+
+    cj.cookies.each do |c|
+      common_cj.add(c)
+    end
+
+    common_cj
+  end
+
+  # Makes a new +HTTP::Cookie+ object with sensible default attributes.
+  #
+  # @return [HTTP::Cookie]
+  def make_default_cookie(cookie_name, cookie_value)
+    HTTP::Cookie.new(
+      {
+        :name => cookie_name,
+        :value => cookie_value,
+        :domain => rhost,
+        :created_at => Time.now,
+        :accessed_at => Time.now,
+        :origin => "http#{ssl ? 's' : ''}://#{rhost}",
+        :secure => ssl, # Are you using HTTPS?
+        :path => '/', # Top level of the domain the cookie applies
+        :for_domain => false, # If true, send cookie to any host in the domain.
+        :httponly => true, # Can the flag be accessed by client side scripts?
+        :expires => nil, # At what dateTime (relevant to the client) should the cookie expire
+        :max_age => nil, # How old should the cookie be when it expires? (Alternative to +:expires+)
+        # If +:expires+ && +:max_age+ are nil, cookie will expire when the current session ends
+      }
+    )
+  end
+


I think we'd be good to move this entirely into the cookie_jar abstraction class directly

I don't think this violates the law of demeter

h00die · 2021-03-29T19:25:32Z

just a note that as a follow on action, the wiki will need to be updated as well.

lib/msf/core/exploit/remote/http/http_cookie.rb

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb

lib/msf/core/exploit/remote/http/http_cookie_jar.rb

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb

lib/msf/core/exploit/remote/http/http_cookie.rb

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb

lib/msf/core/exploit/remote/http/http_cookie_jar.rb

lib/msf/core/exploit/remote/http_client.rb

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb

lib/msf/core/exploit/remote/http/http_cookie_jar.rb

adfoster-r7 · 2021-04-21T18:25:24Z

@agalway-r7 As a pre-requisite to landing this I think we'll want to update the docs to show common workflows for interacting with the cookies/cookie jars etc: https://github.com/rapid7/metasploit-framework/wiki/How-to-Send-an-HTTP-Request-Using-HttpClient

adfoster-r7 · 2021-04-21T18:46:58Z

lib/msf/core/exploit/remote/http_client.rb

+  #
+  # +expire_cookies+ will control if +cleanup+ is called on any passed +Msf::Exploit::Remote::HTTP::HttpCookieJar+ or the client cookiejar


I think this documentation needs updated now that expire_cookies isn't a a parameter

Let's leave this to a follow up PR

adfoster-r7 · 2021-04-26T12:30:10Z

@agalway-r7 Looks like we'll want to update the exchange module too, as I believe it's broken by this pull request - I haven't ran it, just from a static/eyeballing perspective:

metasploit-framework/modules/exploits/windows/http/exchange_ecp_dlp_policy.rb

Line 167 in b177452

unless res.code == 200 && cookie_jar.grep(/^cadata/).any?

Less important, but the spam titan module was recently landed it looks like it uses its own cookie jar as a work around too from what I can see, we could potentially update that - but the above module working is more important.

metasploit-framework/modules/exploits/freebsd/webapp/spamtitan_unauth_rce.rb

Lines 194 to 202 in b177452

    
           # For whatever reason, the web application sometimes returns multiple 
        
           # PHPSESSID cookies and only the last one is valid. So, make sure only 
        
           # the valid one is part of the cookie_jar. 
        
           cookies = res.get_cookies.split(' ') 
        
           php_session = cookies.select { |cookie| cookie.starts_with?('PHPSESSID=') }.last 
        
           cookie_jar.clear 
        
           cookie_jar.add(php_session) 
        
           remaining_cookies = cookies.delete_if { |cookie| cookie.starts_with?('PHPSESSID=') } 
        
           cookie_jar.merge(remaining_cookies)

modules/exploits/windows/http/exchange_ecp_dlp_policy.rb

adfoster-r7 · 2021-04-30T12:41:13Z

lib/msf/core/exploit/remote/http/http_cookie_jar.rb

+          # order.
+          def cookies
+            @cookie_jar.cookies
+          end


We might revisit this API in the future, as asking the cookie jar for its cookies seems a bit weird - as it already is a collection of cookies by definition 🤔

This would pair well with adding support for cookie_jar['cookie_name']

Let's leave this to a follow up PR

adfoster-r7 · 2021-04-30T12:56:18Z

lib/msf/core/exploit/remote/http/http_cookie.rb

+            if domain.nil? || domain.is_a?(DomainName)
+              @cookie.domain = domain
+            else


Suggested change

if domain.nil? || domain.is_a?(DomainName)

@cookie.domain = domain

else

I think this is a leaky internal implementation detail, we should probably delete this in a follow up PR

adfoster-r7 · 2021-04-30T13:27:27Z

Release Notes

Updated the HttpClient mixin with with a new cookie jar implementation which correctly updates and merges the Set-Cookie header responses when using the send_request_cgi keep_cookies option.

adfoster-r7 · 2021-04-30T13:28:50Z

I think this is good to land now 🎉

Follow up work includes:

Add functionality for cookie_jar['cookie_name']
Fix the https/http tests
Removing the leaky DomainName implementation detail
Updating the wiki to show common workflows for interacting with the cookies/cookie jars etc: https://github.com/rapid7/metasploit-framework/wiki/How-to-Send-an-HTTP-Request-Using-HttpClient

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb

h00die · 2021-04-30T18:17:59Z

#14831 (comment)
also, the webapp example module could benefit from having this incorporated

adfoster-r7 · 2021-07-21T16:02:03Z

lib/msf/core/exploit/remote/http_client.rb

    return unless res

    if opts['keep_cookies'] && res.headers['Set-Cookie'].present?
-      # XXX: CGI::Cookie (get_cookies_parsed) is hella broken


Future us: It's hella broken because Set-Cookie becomes comma deliminated, which messes up the parsing logic later down the line:

metasploit-framework/lib/rex/proto/http/packet/header.rb

Lines 49 to 58 in d9e4aad

# Extract each header value pair

header.split(/\r\n/mn).each { |str|

if (md = str.match(/^(.+?)\s*:\s*(.+?)\s*$/))

if (self[md[1]])

self[md[1]] << ", " + md[2]

else

self[md[1]] = md[2]

end

end

}

get_cookies_parsed was only ever tested with a single Set-Cookie header

agalway-r7 added the enhancement label Mar 1, 2021

agalway-r7 changed the title ~~Http cookie jar~~ Add CookieJar support to http_client Mar 1, 2021

adfoster-r7 reviewed Mar 1, 2021

View reviewed changes

modules/exploits/multi/http/gitlab_file_read_rce.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Mar 2, 2021

View reviewed changes

adfoster-r7 marked this pull request as draft March 4, 2021 16:06

adfoster-r7 reviewed Mar 15, 2021

View reviewed changes

agalway-r7 force-pushed the http-cookie-jar branch from 6860b83 to 158bb9b Compare March 24, 2021 16:02

adfoster-r7 mentioned this pull request Mar 31, 2021

Gitea and Gogs Git Hooks RCE #14978

Merged

22 tasks

agalway-r7 force-pushed the http-cookie-jar branch from ff03afc to db8cf61 Compare April 12, 2021 09:53

adfoster-r7 reviewed Apr 12, 2021

View reviewed changes

lib/msf/core/exploit/remote/http/http_cookie.rb Outdated Show resolved Hide resolved

adfoster-r7 reviewed Apr 12, 2021

View reviewed changes

lib/msf/core/exploit/remote/http/http_cookie.rb Show resolved Hide resolved

adfoster-r7 reviewed Apr 12, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb Outdated Show resolved Hide resolved

agalway-r7 force-pushed the http-cookie-jar branch 2 times, most recently from 01d74de to 0afbf17 Compare April 15, 2021 12:39

adfoster-r7 reviewed Apr 15, 2021

View reviewed changes

lib/msf/core/exploit/remote/http/http_cookie_jar.rb Outdated Show resolved Hide resolved

adfoster-r7 reviewed Apr 15, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb Outdated Show resolved Hide resolved

adfoster-r7 reviewed Apr 15, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb Outdated Show resolved Hide resolved

http-client cookie jar support and tests

fc55d74

agalway-r7 force-pushed the http-cookie-jar branch 4 times, most recently from bc5e1da to 3d4ae3b Compare April 16, 2021 15:49

cleanup and removes cookies filtering

88f17c5

agalway-r7 force-pushed the http-cookie-jar branch from 3d4ae3b to 88f17c5 Compare April 16, 2021 16:32

adfoster-r7 reviewed Apr 16, 2021

View reviewed changes

lib/msf/core/exploit/remote/http/http_cookie.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_jar_spec.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb Outdated Show resolved Hide resolved

dwelch-r7 reviewed Apr 19, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb Outdated Show resolved Hide resolved

improvements to tests and api

5df0f0b

adfoster-r7 reviewed Apr 20, 2021

View reviewed changes

consider vhost & expand tests

1b02344

agalway-r7 force-pushed the http-cookie-jar branch from fef5db0 to 1b02344 Compare April 20, 2021 14:13

agalway-r7 marked this pull request as ready for review April 20, 2021 14:41

adfoster-r7 reviewed Apr 21, 2021

View reviewed changes

Update exchange ecp dlp policy to work with new cookie jar

04ff0f6

adfoster-r7 reviewed Apr 29, 2021

View reviewed changes

modules/exploits/windows/http/exchange_ecp_dlp_policy.rb Show resolved Hide resolved

adfoster-r7 reviewed Apr 30, 2021

View reviewed changes

adfoster-r7 merged commit 6c6d769 into rapid7:master Apr 30, 2021

adfoster-r7 mentioned this pull request Apr 30, 2021

Add cookie management to HttpClient and improve standards compliance (Content-Type and PRG pattern) #14139

Merged

adfoster-r7 reviewed Apr 30, 2021

View reviewed changes

spec/lib/msf/core/exploit/remote/remote/http/http_cookie_spec.rb Show resolved Hide resolved

This was referenced May 4, 2021

Update cacti rce to use cookie jar api h00die/metasploit-framework#17

Merged

Ensure cookie jars are correctly duped #15149

Merged

ebleiweiss-r7 added the rn-enhancement release notes enhancement label May 6, 2021

agalway-r7 mentioned this pull request May 7, 2021

Converts Cockpit CMS to use new CookieJar API #15164

Merged

9 tasks

adfoster-r7 reviewed Jul 21, 2021

View reviewed changes

		common_cj = HTTP::CookieJar.new
		cookie_jar.each { \|c\| common_cj.add(c) }

	self.cookie_jar = HTTP::CookieJar.new
	self.cookie_jar = Exploit::Remote::HttpCookieJar.new

		#
		# +expire_cookies+ will control if +cleanup+ is called on any passed +Msf::Exploit::Remote::HTTP::HttpCookieJar+ or the client cookiejar

	if domain.nil? \|\| domain.is_a?(DomainName)
	@cookie.domain = domain
	else

	# Extract each header value pair
	header.split(/\r\n/mn).each { \|str\|
	if (md = str.match(/^(.+?)\s:\s(.+?)\s*$/))
	if (self[md[1]])
	self[md[1]] << ", " + md[2]
	else
	self[md[1]] = md[2]
	end
	end
	}

Conversation

agalway-r7 commented Mar 1, 2021 • edited by wvu Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Verification

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

adfoster-r7 Mar 4, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

agalway-r7 Mar 2, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

adfoster-r7 commented Mar 2, 2021

Uh oh!

agalway-r7 commented Mar 2, 2021

Uh oh!

wvu commented Mar 4, 2021

Uh oh!

agalway-r7 commented Mar 9, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

h00die commented Mar 29, 2021

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

adfoster-r7 commented Apr 21, 2021

Uh oh!

agalway-r7 commented Mar 1, 2021 •

edited by wvu

Loading

adfoster-r7 Mar 4, 2021 •

edited

Loading

agalway-r7 Mar 2, 2021 •

edited

Loading

agalway-r7 commented Mar 9, 2021 •

edited

Loading

adfoster-r7 commented Apr 30, 2021 •

edited by pbarry-r7

Loading

adfoster-r7 commented Apr 30, 2021 •

edited

Loading