po/todo: first pass of robustness tests, added todo for remove/rename

[git.ikiwiki.info.git] / doc / plugins / htmlscrubber.mdwn

diff --git a/doc/plugins/htmlscrubber.mdwn b/doc/plugins/htmlscrubber.mdwn
index 8e37a5bd5e9d536d105aabef7d1aaf203ddba4e6..7db372e1be272b185dc735f8fb27a79ed612420c 100644 (file)
--- a/doc/plugins/htmlscrubber.mdwn
+++ b/doc/plugins/htmlscrubber.mdwn
@@ -1,5 +1,5 @@
-[[template id=plugin name=htmlscrubber core=1 author="[[Joey]]"]]
-[[tag type/html]]
+[[!template id=plugin name=htmlscrubber core=1 author="[[Joey]]"]]
+[[!tag type/html]]
 
 This plugin is enabled by default. It sanitizes the html on pages it renders
 to avoid XSS attacks and the like.
@@ -7,9 +7,13 @@ to avoid XSS attacks and the like.
 It excludes all html tags and attributes except for those that are
 whitelisted using the same lists as used by Mark Pilgrim's Universal Feed
 Parser, documented at <http://feedparser.org/docs/html-sanitization.html>.
-Notably it strips `style`, `link`, and the `style` attribute.
+Notably it strips `style` and `link` tags, and the `style` attribute.
 
-It uses the [[cpan HTML::Scrubber]] perl module to perform its html
+All attributes that can be used to specify an url are checked to make sure
+that the url is in a known, safe scheme, and to block embedded javascript
+in such urls.
+
+It uses the [[!cpan HTML::Scrubber]] perl module to perform its html
 sanitisation, and this perl module also deals with various entity encoding
 tricks.
 
@@ -17,17 +21,29 @@ While I believe that this makes ikiwiki as resistant to malicious html
 content as anything else on the web, I cannot guarantee that it will
 actually protect every user of every browser from every browser security
 hole, badly designed feature, etc. I can provide NO WARRANTY, like it says
-in ikiwiki's [GPL](GPL) license. 
+in ikiwiki's [[GPL]] license. 
 
 The web's security model is *fundamentally broken*; ikiwiki's html
 sanitisation is only a patch on the underlying gaping hole that is your web
 browser.
 
+Note that enabling or disabling the htmlscrubber plugin also affects some
+other HTML-related functionality, such as whether [[meta]] allows
+potentially unsafe HTML tags.
+
+The `htmlscrubber_skip` configuration setting can be used to skip scrubbing
+of some pages. Set it to a [[PageSpec]], such as "!*/Discussion", and pages
+matching that can have all the evil CSS, JavsScript, and unsafe html
+elements you like. One safe way to use this is to use [[lockedit]] to lock
+those pages, so only admins can edit them.
+
 ----
 
 Some examples of embedded javascript that won't be let through when this
 plugin is active:
 
-* <span style="background: url(javascript:window.location='http://example.org/')">test</span>
-* <span style="&#x61;&#x6e;&#x79;&#x3a;&#x20;&#x65;&#x78;&#x70;&#x72;&#x65;&#x73;&#x73;&#x69;&#x6f;&#x6e;&#x28;&#x77;&#x69;&#x6e;&#x64;&#x6f;&#x77;&#x2e;&#x6c;&#x6f;&#x63;&#x61;&#x74;&#x69;&#x6f;&#x6e;&#x3d;&#x27;&#x68;&#x74;&#x74;&#x70;&#x3a;&#x2f;&#x2f;&#x65;&#x78;&#x61;&#x6d;&#x70;&#x6c;&#x65;&#x2e;&#x6f;&#x72;&#x67;&#x2f;&#x27;&#x29;">test</span>
-* <span style="&#97;&#110;&#121;&#58;&#32;&#101;&#120;&#112;&#114;&#101;&#115;&#115;&#105;&#111;&#110;&#40;&#119;&#105;&#110;&#100;&#111;&#119;&#46;&#108;&#111;&#99;&#97;&#116;&#105;&#111;&#110;&#61;&#39;&#104;&#116;&#116;&#112;&#58;&#47;&#47;&#101;&#120;&#97;&#109;&#112;&#108;&#101;&#46;&#111;&#114;&#103;&#47;&#39;&#41;">test</span>
+* script tag test <script>window.location='http://example.org';</script>
+* <span style="background: url(javascript:window.location='http://example.org/')">CSS script test</span>
+* <span style="&#x61;&#x6e;&#x79;&#x3a;&#x20;&#x65;&#x78;&#x70;&#x72;&#x65;&#x73;&#x73;&#x69;&#x6f;&#x6e;&#x28;&#x77;&#x69;&#x6e;&#x64;&#x6f;&#x77;&#x2e;&#x6c;&#x6f;&#x63;&#x61;&#x74;&#x69;&#x6f;&#x6e;&#x3d;&#x27;&#x68;&#x74;&#x74;&#x70;&#x3a;&#x2f;&#x2f;&#x65;&#x78;&#x61;&#x6d;&#x70;&#x6c;&#x65;&#x2e;&#x6f;&#x72;&#x67;&#x2f;&#x27;&#x29;">entity-encoded CSS script test</span>
+* <span style="&#97;&#110;&#121;&#58;&#32;&#101;&#120;&#112;&#114;&#101;&#115;&#115;&#105;&#111;&#110;&#40;&#119;&#105;&#110;&#100;&#111;&#119;&#46;&#108;&#111;&#99;&#97;&#116;&#105;&#111;&#110;&#61;&#39;&#104;&#116;&#116;&#112;&#58;&#47;&#47;&#101;&#120;&#97;&#109;&#112;&#108;&#101;&#46;&#111;&#114;&#103;&#47;&#39;&#41;">entity-encoded CSS script test</span>
+* <a href="javascript&#x3A;alert('foo')">click me</a>