🔒 Improved validation of fetched urls and responses in oembed endpoint

no issue

- prevent oembed fetching from accessing IP addresses or localhost domains
- prevent oembed endpoint from passing through fetched responses as-is
  - reject any fetched data that does not validate against the oembed spec
  - strip any unknown properties from the oembed response before returning

Credits: Nick Mykhailyshyn
This commit is contained in:
Kevin Ansfield 2020-03-03 12:47:18 +00:00
parent afb3177e5e
commit 4773939670
2 changed files with 361 additions and 11 deletions

View File

@ -3,6 +3,7 @@ const {extract, hasProvider} = require('oembed-parser');
const Promise = require('bluebird');
const request = require('../../lib/request');
const cheerio = require('cheerio');
const _ = require('lodash');
async function fetchBookmarkData(url, html) {
const metascraper = require('metascraper')([
@ -74,10 +75,6 @@ const findUrlWithProvider = (url) => {
return {url, provider};
};
const getOembedUrlFromHTML = (html) => {
return cheerio('link[type="application/json+oembed"]', html).attr('href');
};
function unknownProvider(url) {
return Promise.reject(new common.errors.ValidationError({
message: common.i18n.t('errors.api.oembed.unknownProvider'),
@ -93,12 +90,40 @@ function knownProvider(url) {
});
}
function fetchOembedData(url) {
let provider;
({url, provider} = findUrlWithProvider(url));
function isIpOrLocalhost(url) {
try {
const IPV4_REGEX = /^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$/;
const IPV6_REGEX = /:/; // fqdns will not have colons
const HTTP_REGEX = /^https?:/i;
let {protocol, hostname} = new URL(url);
if (!HTTP_REGEX.test(protocol) || hostname === 'localhost' || IPV4_REGEX.test(hostname) || IPV6_REGEX.test(hostname)) {
return true;
}
return false;
} catch (e) {
return true;
}
}
function fetchOembedData(_url) {
// parse the url then validate the protocol and host to make sure it's
// http(s) and not an IP address or localhost to avoid potential access to
// internal network endpoints
if (isIpOrLocalhost(_url)) {
return unknownProvider();
}
// check against known oembed list
let {url, provider} = findUrlWithProvider(_url);
if (provider) {
return knownProvider(url);
}
// url not in oembed list so fetch it in case it's a redirect or has a
// <link rel="alternate" type="application/json+oembed"> element
return request(url, {
method: 'GET',
timeout: 2 * 1000,
@ -107,19 +132,70 @@ function fetchOembedData(url) {
'user-agent': 'Ghost(https://github.com/TryGhost/Ghost)'
}
}).then((response) => {
// url changed after fetch, see if we were redirected to a known oembed
if (response.url !== url) {
({url, provider} = findUrlWithProvider(response.url));
if (provider) {
return knownProvider(url);
}
}
if (provider) {
return knownProvider(url);
// check for <link rel="alternate" type="application/json+oembed"> element
let oembedUrl;
try {
oembedUrl = cheerio('link[type="application/json+oembed"]', response.body).attr('href');
} catch (e) {
return unknownProvider(url);
}
const oembedUrl = getOembedUrlFromHTML(response.body);
if (oembedUrl) {
// make sure the linked url is not an ip address or localhost
if (isIpOrLocalhost(oembedUrl)) {
return unknownProvider(oembedUrl);
}
// fetch oembed response from embedded rel="alternate" url
return request(oembedUrl, {
method: 'GET',
json: true
}).then((response) => {
return response.body;
// validate the fetched json against the oembed spec to avoid
// leaking non-oembed responses
const body = response.body;
const hasRequiredFields = body.type && body.version;
const hasValidType = ['photo', 'video', 'link', 'rich'].includes(body.type);
if (hasRequiredFields && hasValidType) {
// extract known oembed fields from the response to limit leaking of unrecognised data
const knownFields = [
'type',
'version',
'html',
'url',
'title',
'width',
'height',
'author_name',
'author_url',
'provider_name',
'provider_url',
'thumbnail_url',
'thumbnail_width',
'thumbnail_height'
];
const oembed = _.pick(body, knownFields);
// ensure we have required data for certain types
if (oembed.type === 'photo' && !oembed.url) {
return;
}
if ((oembed.type === 'video' || oembed.type === 'rich') && (!oembed.html || !oembed.width || !oembed.height)) {
return;
}
// return the extracted object, don't pass through the response body
return oembed;
}
}).catch(() => {});
}
});

View File

@ -56,4 +56,278 @@ describe('Oembed API', function () {
done();
});
});
describe('with unknown provider', function () {
it('fetches url and follows <link rel="alternate">', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://test.com/oembed"></head></html>');
const oembedMock = nock('http://test.com')
.get('/oembed')
.reply(200, {
version: '1.0',
type: 'link'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(200)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.true();
done();
});
});
it('rejects invalid oembed responses', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://test.com/oembed"></head></html>');
const oembedMock = nock('http://test.com')
.get('/oembed')
.reply(200, {
version: '1.0',
html: 'test'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.true();
done();
});
});
it('rejects unknown oembed types', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://test.com/oembed"></head></html>');
const oembedMock = nock('http://test.com')
.get('/oembed')
.reply(200, {
version: '1.0',
type: 'unknown'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.true();
done();
});
});
it('rejects invalid photo responses', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://test.com/oembed"></head></html>');
const oembedMock = nock('http://test.com')
.get('/oembed')
.reply(200, {
// no `url` field
version: '1.0',
type: 'photo',
thumbnail_url: 'https://test.com/thumbnail.jpg'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.true();
done();
});
});
it('rejects invalid video responses', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://test.com/oembed"></head></html>');
const oembedMock = nock('http://test.com')
.get('/oembed')
.reply(200, {
// no `html` field
version: '1.0',
type: 'video',
thumbnail_url: 'https://test.com/thumbnail.jpg'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.true();
done();
});
});
it('strips unknown response fields', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://test.com/oembed"></head></html>');
const oembedMock = nock('http://test.com')
.get('/oembed')
.reply(200, {
version: '1.0',
type: 'video',
html: '<p>Test</p>',
width: 200,
height: 100,
unknown: 'test'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(200)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.true();
res.body.should.deepEqual({
version: '1.0',
type: 'video',
html: '<p>Test</p>',
width: 200,
height: 100
});
should.not.exist(res.body.unknown);
done();
});
});
it('skips fetching IPv4 addresses', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://192.168.0.1/oembed"></head></html>');
const oembedMock = nock('http://192.168.0.1')
.get('/oembed')
.reply(200, {
version: '1.0',
type: 'link'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.false();
done();
});
});
it('skips fetching IPv6 addresses', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://[2607:f0d0:1002:51::4]:9999/oembed"></head></html>');
const oembedMock = nock('http://[2607:f0d0:1002:51::4]:9999')
.get('/oembed')
.reply(200, {
version: '1.0',
type: 'link'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.false();
done();
});
});
it('skips fetching localhost', function (done) {
const pageMock = nock('http://test.com')
.get('/')
.reply(200, '<html><head><link rel="alternate" type="application/json+oembed" href="http://localhost:9999/oembed"></head></html>');
const oembedMock = nock('http://localhost:9999')
.get('/oembed')
.reply(200, {
// no `html` field
version: '1.0',
type: 'video',
thumbnail_url: 'https://test.com/thumbnail.jpg'
});
const url = encodeURIComponent('http://test.com');
request.get(localUtils.API.getApiQuery(`oembed/?url=${url}`))
.set('Origin', config.get('url'))
.expect('Content-Type', /json/)
.expect('Cache-Control', testUtils.cacheRules.private)
.expect(422)
.end(function (err, res) {
if (err) {
return done(err);
}
pageMock.isDone().should.be.true();
oembedMock.isDone().should.be.false();
done();
});
});
});
});