WP Content Security Plugin

When you first turn on CSP, put into report-only mode and build the basic rules for your site. After about a week, turn off report-only and go to enforce rules. If you want to implement the latest W3C version of CSP - version 3 Google recommends - set the following for default-src, script-src, and style-src: 'unsafe-inline' 'unsafe-eval' 'strict-dynamic' https: http: (single quotes are required) This will allow modern browsers to run the latest version of CSP with nonces, etc. and older browsers to just work without restrictions. If you're going to run CSP v2, one good way of building a policy for a site would be to begin with a default-src of 'self', and to build up a policy from there that contains only those resource types which are actually in use for the page you'd like to protect. If you don't use webfonts, for instance, there's no reason to specify a source list for font-src; specifying only those resource types a page uses ensures that the possible attack surface for that page remains as small as possible.

Usually you will trust your own site for all directives; however, I usually only add 'self' when it shows up as a violation. None of these directives is inherited, except some directives will default to 'default-src' if not set explicitly.

Usually you would want to keep security as strict as possible while still allowing your application to run. Therefore, '*' should be avoided.

The W3C specification allows for a different policy for each page, this plugin was not written with page-level security capability.

The W3C specification allows for this functionality; this plugin does not support this capability.

1. First check that your site is producing CSP errors by starting the developer tools in your browser (usually F12) and checking for messages in the console output.
2. If nothing is in the console output then check the page has a CSP header by looking at the page in the 'network' tab of the dev tools. Check the 'response' has a header called 'content-security-policy' or 'content-security-policy-report-only' - if this is misisng then the plugin is not running or CSP is not enabled.
3. If there is a CSP header and nothing is reported in the console then you have no violations and everything is running as it should. Yippee!
4. If there is a CSP header and errors in the console then the REST route might not be registered properly. Go to /wp-json and look for 'wpcsp' (usually CTRL-F for find and type in wpcsp) - if nothing is listed then the REST route is not getting registered.
5. Look in the PHP error logs for an error - post the error, file name and line number in the support forums and I should be able to work out why it's failing.

Inline scripts and styles can be dangerous, you do not know which scripts wrote them and probably don't want them run if you can avoid it. When you use 'script-dynamic', the "unsafe-eval" and "unsafe-inline stop working and the browser will say in the console "Note that 'unsafe-inline' is ignored if either a hash or nonce value is present in the source list." To fix this either:

• Put all the scripts and/or style code into files and include the files.
• If the browser returns "Either the 'unsafe-inline' keyword, a hash ('sha256-h3SEZNZpOYg4jp6TCkoWN7Z477Qt3q1owH0SPbz+a4M='), or a nonce ('nonce-...') is required to enable inline execution." - you can take the SHA number (including single quotes) and put that in the policy line.

This is different for all sites. The plugin will automatically delete records older than one week to keep the size managable. Also, if too many records are found the system will only report on the worse errors to avoid locking your browser.

Every error output by your browser is likely to result in a call to the server to log the error - if a page has 20 errors that's 20 calls to your server - this can be a lot of processing power. To avoid this change the "Log Violations" option from "Yes, All" to "Yes - 10%", "Yes - 1%', or "Yes - 0.1%" - in each case the plugin will randomly allow only a set fraction of your visitors to report errors back to the server, they're still enfored at the browser but no report will come back to your site.