If you want to set security config open for a certain path, you can configure match option.
For example, just open csp when path contains /example, you can configure with the following configuration:
exports.security = {
csp: {
match: '/example',
// match: /^\/api/, // support regexp
// match: ctx => ctx.path.startsWith('/api'), // support function
// match: [ ctx => ctx.path.startsWith('/api'), /^\/foo$/, '/bar'], // support Array
policy: {
//...
},
},
};
If you want to set security config disable for a certain path, you can configure ignore option.
For example, just disable xframe when path contains /example while our pages can be embedded in cooperative businesses , you can configure with the following configuration:
exports.security = {
xframe: {
ignore: '/example',
// ignore: /^\/api/, // support regexp
// ignore: ctx => ctx.path.startsWith('/api'), // support function
// ignore: [ ctx => ctx.path.startsWith('/api'), /^\/foo$/, '/bar'], // support Array
// ...
},
};
mention:match has higher priority than ignore
Dynamic configuration for security plugins depend on context
There are times when we want to be more flexible to configure security plugins.For example:
To decide whether to enable or disable the xframe security header from the context of the request.
To decide csp policies from different request urls.
Then we can configure ctx.securityOptions[name] opts in the custom middleware or controller,then the current request configuration will overrides the default configuration (new configuration will be merged and override the default project configuration, but only take effect in the current request)
Not all security plugins support dynamic configuration, only following plugins list support
csp
hsts
noopen
nosniff
xframe
xssProtection
And in helper:
shtml
helper is the same way to configure.
ctx.securityOptions.shtml = {
whiteList: {
},
};
Mention
Security is a big thing, please pay attention to the risk of changes in the security configuration (especially dynamic changes)
ctx.securityOptions the current request configuration will overrides the default configuration, but it does not make a deep copy,so pay attention to configure csp.policy, it will not be merged.
If you configure ctx.securityOptions,please write unit tests to ensure the code is correct.
API
ctx.isSafeDomain(domain)
Whether or not the domain is in the whitelist of the configuration. See ctx.redirect.
Note: egg-cors module uses this function internally to determine whether or not send back an Access-Control-Allow-Origin response header with the value of safe domain. Otherwise, ignore the request with an error, No 'Access-Control-Allow-Origin' header is present on the requested resource.
Generally used when send POST form request. When page rendering, put ctx.csrf into form hidden field or query string.(_csrf is the key).
When submitting the form, please submit with the _csrf token parameter.
CSRF token will also set to cookie by default, and you can send token through header:
In jQuery:
var csrftoken = Cookies.get('csrftoken');
function csrfSafeMethod(method) {
// these HTTP methods do not require CSRF protection
return (/^(GET|HEAD|OPTIONS|TRACE)$/.test(method));
}
$.ajaxSetup({
beforeSend: function(xhr, settings) {
if (!csrfSafeMethod(settings.type) && !this.crossDomain) {
xhr.setRequestHeader('x-csrf-token', csrftoken);
}
},
});
Options
there are some options that you can customize:
exports.security = {
csrf: {
type: 'ctoken', // can be ctoken, referer, all or any, default to ctoken
useSession: false, // if useSession set to true, the secret will keep in session instead of cookie
ignoreJSON: false, // skip check JSON requests if ignoreJSON set to true
cookieName: 'csrfToken', // csrf token's cookie name
sessionName: 'csrfToken', // csrf token's session name
headerName: 'x-csrf-token', // request csrf token's name in header
bodyName: '_csrf', // request csrf token's name in body
queryName: '_csrf', // request csrf token's name in query
refererWhiteList: [], // referer white list
},
}
Rotate CSRF secret
Must call ctx.rotateCsrfSecret() when user login to ensure each user has independent secret.
safe redirect
ctx.redirect(url) If url is not in the configuration of the white list, the redirect will be prohibited
ctx.unsafeRedirect(url) Not Recommended;
Security plugin override ctx.redirect method,all redirects will be judged by the domain name.
If you need to use ctx.redirect, you need to do the following configuration in the application configuration file:
If user do not configure domainWhiteList or domainWhiteList is empty, it will pass all redirects, equal to ctx.unsafeRedirect(url). domainWhiteList and url are case insensitive.
Used to output variables in javascript(include onload/event),it will do JAVASCRIPT ENCODE for the variable string.It will escape all characters to \x which are not in the whitelist to avoid XSS attack.
const foo = '"hello"';
// not use sjs
console.log(`var foo = "${foo}";`);
// => var foo = ""hello"";
// use sjs
console.log(`var foo = "${ctx.helper.sjs(foo)}";`);
// => var foo = "\\x22hello\\x22";
.shtml()
If you want to output richtexts in views, you need to use shtml helper.
It will do XSS filter, then output html tags to avoid illegal scripts.
** shtml is a very complex process, it will effect server performance, so if you do not need to output HTML, please do not use shtml.**
Examples:
// js
const value = `<a href="http://www.domain.com">google</a><script>evilcode…</script>`;
// in your view
<html>
<body>
${helper.shtml($value)}
</body>
</html>
// => <a href="http://www.domain.com">google</a><script>evilcode…</script>
shtml based on xss, and add filter by domain feature.
For example, only support a tag, and filter all attributes except for title:
whiteList: {a: ['title']}
options:
config.helper.shtml.domainWhiteList has been deprecated, please use config.security.domainWhiteList instead.
Mention that shtml uses a strict white list mechanism, in addition to filtering out the XSS risk of the string,tags and attrs which are not in the default rule will be filtered.
For example html tag is not in the whitelist.
const html = '<html></html>';
// html
${helper.shtml($html)}
// output none
Commonly used data-xx property is not in the whitelist, so it will be filtered.
So please check the applicable scenarios for shtml, it usually used for richtext submmited by user.
A usage error will limit functions, also affect the performance of the server.
Such scenes are generally forums, comments, etc.
Even if the forum does not support the HTML content input, do not use this helper, you can directly use escape instead.
.spath()
If you want to use users input for a file path, please use spath for security check. If path is illegal, it will return null.
Illegal path:
relative path starts with ..
absolute path starts with /
above path try to use url encode to bypass the check
If you want to output json in javascript without encoding, it will be a risk for XSS.
sjson supports json encode,it will iterate all keys in json, then escape all characters in the value to \x to avoid XSS attack, and keep the json structure unchanged.
If you want to output json string in your views, please use ${ctx.helper.sjson(var)}to escape.
it has a very complex process and will lost performance, so avoid the use as far as possible
It will cause remote command execution vulnerability, when user submit the implementation of the command by browser.because the server does not filter for the implementation of the function, resulting in the execution of the command can usually lead to the invasion of the server.
If you want to get user submit for command's parameter, please use cliFilter。
Escape command line arguments. Add single quotes around a string and quotes/escapes any existing single quotes allowing you to pass a string directly to a shell function and having it be treated as a single safe argument.
Disabled by default. If your website based on https, we recommend you should enable it.
maxAge one year by default 365 * 24 * 3600
includeSubdomains false by default
csp
Default disabled. If you need to enable, please contact your security engineers and determine the opening strategy
policy policies used by csp
X-Download-Options:noopen
Default enabled, disable IE download dialog automatically open download file and will cause XSS
X-Content-Type-Options:nosniff
Disable IE8's auto MIME sniffing. E.g: take text/plain as text/html by mistake and render it, especially when there's something untrusted in the local service.
X-Frame-Options
Defaulting to "SAMEORIGIN", only allow iframe embed by same origin.
value Defaulting to SAMEORIGIN
X-XSS-Protection
disable Defaulting to false,same as 1; mode=block.
SSRF Protection
In a Server-Side Request Forgery (SSRF) attack, the attacker can abuse functionality on the server to read or update internal resources.
egg-security provide ctx.safeCurl, app.safeCurl and agent.safeCurl to provide http request(like ctx.curl, app.curl and agent.curl) with SSRF protection.
Configuration
ipBlackList(Array) - specific which ip are illegal when request with safeCurl.
ipExceptionList(Array) - specific which ip are legal within ipBlackList.
checkAddress(Function) - determine the ip by the function's return value, false means illegal ip.
// config/config.default.js
exports.security = {
ssrf: {
// support both cidr subnet or specific ip
ipBlackList: [
'10.0.0.0/8',
'127.0.0.1',
'0.0.0.0/32',
],
// support both cidr subnet or specific ip
ipExceptionList: [
'10.1.1.1',
'10.10.0.1/24',
],
// checkAddress has higher priority than ipBlackList
checkAddress(ip) {
return ip !== '127.0.0.1';
}
},
};