security considerations document
This commit is contained in:
parent
1fb5f15ead
commit
3fd7e48907
|
@ -1,6 +1,5 @@
|
||||||
|
#!/bin/ruby
|
||||||
|
# frozen_string_literal: true
|
||||||
|
|
||||||
|
require 'optionparser'
|
||||||
|
|
||||||
if __FILE__ == $0
|
|
||||||
text = $stdin.read
|
|
||||||
renderer = MDPP::Renderer.new(text, {})
|
|
||||||
puts renderer.render
|
|
||||||
end
|
|
||||||
|
|
|
@ -0,0 +1,21 @@
|
||||||
|
Security acknowledgements
|
||||||
|
=========================
|
||||||
|
|
||||||
|
While special care has been taken to prevent some of the more common common
|
||||||
|
vulnerabilities that might arise from using this parser, it does not prevent
|
||||||
|
certain issues which **which should be acknowledged**.
|
||||||
|
|
||||||
|
- It is possible to inject a form of one-click XSS into the website. In
|
||||||
|
particular, there are no restrictions placed on urls embedded within the links
|
||||||
|
(as per the description of CommonMark specification). As such, something as
|
||||||
|
simple as `[test](<javascript:dangerous code here>)` would be more than enough
|
||||||
|
to employ such an exploit.
|
||||||
|
- While generally speaking the parser acts stable on most tests, and precents
|
||||||
|
stray HTML tokens from occuring in the output text where appropriate, due to
|
||||||
|
the nontrivial nature of the task some form of XSS injection may or may not
|
||||||
|
occur. If such an incident occurs, please report it to the current maintainer
|
||||||
|
of the project.
|
||||||
|
- User input should NOT be trusted when it comes to applying options to
|
||||||
|
rendering. Some renderers, such as the HTML renderer, allow modifying the
|
||||||
|
style parameter for rendered tags, which when passed control of to an
|
||||||
|
untrusted party may become an XSS attack vector.
|
Loading…
Reference in New Issue