Reflected XSS in GraphQL Playground

説明

Impact

directly impacted:

  • graphql-playground-html@<1.6.22 - all unsanitized user input for renderPlaygroundPage()

all of our consuming packages of graphql-playground-html are impacted:

  • graphql-playground-middleware-express@<1.7.16 - unsanitized user input to expressPlayground()
  • graphql-playground-middleware-koa@<1.6.15 - unsanitized user input to koaPlayground()
  • graphql-playground-middleware-lambda@<1.7.17 - unsanitized user input to lambdaPlayground()
  • graphql-playground-middleware-hapi@<1.6.13 - unsanitized user input to hapiPlayground()

as well as any other packages that use these methods with unsanitized user input.

not impacted:

  • graphql-playground-electron - uses renderPlaygroundPage() statically for a webpack build for electron bundle, no dynamic user input
  • graphql-playground-react - usage of the component directly in a react application does not expose reflected XSS vulnerabilities. only the demo in public/ contains the vulnerability, because it uses an old version of the html pacakge.

Patches

upgrading to the above mentioned versions will solve the issue.

If you're using graphql-playground-html directly, then:

yarn add graphql-playground-html@^1.6.22

or

npm install --save graphql-playground-html@^1.6.22

Then, similar steps need to be taken for each middleware:

Workarounds

Ensure you properly sanitize all user input for options you use for whatever function to initialize GraphQLPlayground:

for example, with graphql-playground-html and express:

const { sanitizeUrl } = require('@braintree/sanitize-url');

const qs = require('querystringify');

const { renderPlaygroundPage } = require('graphql-playground-html');

module.exports = (req, res, next) => {
    const { endpoint } = qs.parse(req.url)
    res.html(renderPlaygroundPage({endpoint: sanitizeUrl(endpoint) })).status(200)
    next()
}

or, with graphql-playground-express:

const { expressPlayground } = require('graphql-playground-middleware-express');
const { sanitizeUrl } = require('@braintree/sanitize-url');

const qs = require('querystringify');

const { renderPlaygroundPage } = require('graphql-playground-html');

module.exports = (req, res, next) => {
    const { endpoint } = qs.parse(req.url)
    res.html(expressPlayground({endpoint: sanitizeUrl(endpoint) })).status(200)
    next()
}

References

Credits

Masato Kinugawa of Cure53

For more information

If you have any questions or comments about this advisory:
* Open an issue in graphql-playground
* Email us at [email protected]

基本情報

タイプ
reviewed
深刻度
high
GitHub 上のアドバイザリ
アドバイザリを開く ↗
リポジトリのアドバイザリ
リポジトリのアドバイザリを開く ↗
ソースコード
未指定
公開(アドバイザリ)
2020-06-09 00:24:57 UTC
更新
2023-01-09 05:02:21 UTC
GitHub レビュー済み
2020-06-08 20:27:53 UTC

EPSS Score

Score Percentile
33.49% 96.84%

CVSS Scores

Base score Version Severity Vector
7.4 3.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:N/I:H/A:N クリックして展開
攻撃ベクター (AV:N)
インターネットなど、ルーティングされたネットワーク越しに遠隔から悪用しうる。端末の前にいる必要はない。
攻撃の複雑さ (AC:L)
攻撃者が条件を満たせば、レース条件や珍しい構成に依存せずに再現しやすい。
必要な権限 (PR:N)
事前のログインや昇格は不要で、匿名アクセスのまま踏み台にしうる。
ユーザーの関与 (UI:R)
インストールの許可、設定変更、悪意あるファイルの実行など、人の一度の判断がトリガーになる。
スコープ (S:C)
脆弱箇所を足がかりに、別コンポーネントや別権限域まで影響が広がりうる。
機密性への影響 (C:N)
機微情報の漏えいは想定しにくい。
完全性への影響 (I:H)
権限の奪取や広範なログ改竄など、システムの信頼根拠を揺るがす改ざんが現実的。
可用性への影響 (A:N)
業務継続に支障が出るレベルの停止や劣化は想定されない。

Identifiers

CWEs

CWE id Name
CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')

Affected packages (1)

Vulnerable version ranges and first patched releases as published by GitHub.

Ecosystem Package Vulnerable range First patched Vulnerable functions
npm graphql-playground-html < 1.6.22 1.6.22

References

cvelogic Threat Intelligence