feat: introduce new env ALLOW_UNSAFE_DATA_SCHEME to allow rendering data uri scheme (#21321)

docker/.env.example

@@ -826,6 +826,9 @@ MAX_ITERATIONS_NUM=99
 # The timeout for the text generation in millisecond
 TEXT_GENERATION_TIMEOUT_MS=60000
 
+# Allow rendering unsafe URLs which have "data:" scheme.
+ALLOW_UNSAFE_DATA_SCHEME=false
+
 # ------------------------------
 # Environment Variables for db Service
 # ------------------------------

docker/docker-compose-template.yaml

@@ -67,6 +67,7 @@ services:
       TEXT_GENERATION_TIMEOUT_MS: ${TEXT_GENERATION_TIMEOUT_MS:-60000}
       CSP_WHITELIST: ${CSP_WHITELIST:-}
       ALLOW_EMBED: ${ALLOW_EMBED:-false}
+      ALLOW_UNSAFE_DATA_SCHEME: ${ALLOW_UNSAFE_DATA_SCHEME:-false}
       MARKETPLACE_API_URL: ${MARKETPLACE_API_URL:-https://marketplace.dify.ai}
       MARKETPLACE_URL: ${MARKETPLACE_URL:-https://marketplace.dify.ai}
       TOP_K_MAX_VALUE: ${TOP_K_MAX_VALUE:-}

docker/docker-compose.yaml

@@ -364,6 +364,7 @@ x-shared-env: &shared-api-worker-env
   MAX_PARALLEL_LIMIT: ${MAX_PARALLEL_LIMIT:-10}
   MAX_ITERATIONS_NUM: ${MAX_ITERATIONS_NUM:-99}
   TEXT_GENERATION_TIMEOUT_MS: ${TEXT_GENERATION_TIMEOUT_MS:-60000}
+  ALLOW_UNSAFE_DATA_SCHEME: ${ALLOW_UNSAFE_DATA_SCHEME:-false}
   POSTGRES_USER: ${POSTGRES_USER:-${DB_USERNAME}}
   POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-${DB_PASSWORD}}
   POSTGRES_DB: ${POSTGRES_DB:-${DB_DATABASE}}
@@ -582,6 +583,7 @@ services:
       TEXT_GENERATION_TIMEOUT_MS: ${TEXT_GENERATION_TIMEOUT_MS:-60000}
       CSP_WHITELIST: ${CSP_WHITELIST:-}
       ALLOW_EMBED: ${ALLOW_EMBED:-false}
+      ALLOW_UNSAFE_DATA_SCHEME: ${ALLOW_UNSAFE_DATA_SCHEME:-false}
       MARKETPLACE_API_URL: ${MARKETPLACE_API_URL:-https://marketplace.dify.ai}
       MARKETPLACE_URL: ${MARKETPLACE_URL:-https://marketplace.dify.ai}
       TOP_K_MAX_VALUE: ${TOP_K_MAX_VALUE:-}

web/.env.example

@@ -32,6 +32,9 @@ NEXT_PUBLIC_CSP_WHITELIST=
 # Default is not allow to embed into iframe to prevent Clickjacking: https://owasp.org/www-community/attacks/Clickjacking
 NEXT_PUBLIC_ALLOW_EMBED=
 
+# Allow rendering unsafe URLs which have "data:" scheme.
+NEXT_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME=false
+
 # Github Access Token, used for invoking Github API
 NEXT_PUBLIC_GITHUB_ACCESS_TOKEN=
 # The maximum number of top-k value for RAG.

web/app/components/base/markdown-blocks/utils.ts

@@ -1,3 +1,7 @@
+import { ALLOW_UNSAFE_DATA_SCHEME } from '@/config'
+
 export const isValidUrl = (url: string): boolean => {
-  return ['http:', 'https:', '//', 'mailto:'].some(prefix => url.startsWith(prefix))
+  const validPrefixes = ['http:', 'https:', '//', 'mailto:']
+  if (ALLOW_UNSAFE_DATA_SCHEME) validPrefixes.push('data:')
+  return validPrefixes.some(prefix => url.startsWith(prefix))
 }

web/app/components/base/markdown/markdown-utils.ts

@@ -4,6 +4,7 @@
  * Includes preprocessing for LaTeX and custom "think" tags.
  */
 import { flow } from 'lodash-es'
+import { ALLOW_UNSAFE_DATA_SCHEME } from '@/config'
 
 export const preprocessLaTeX = (content: string) => {
   if (typeof content !== 'string')
@@ -86,5 +87,8 @@ export const customUrlTransform = (uri: string): string | undefined => {
   if (PERMITTED_SCHEME_REGEX.test(scheme))
     return uri
 
+  if (ALLOW_UNSAFE_DATA_SCHEME && scheme === 'data:')
+    return uri
+
   return undefined
 }

web/app/layout.tsx

@@ -44,6 +44,7 @@ const LocaleLayout = async ({
     [DatasetAttr.DATA_PUBLIC_LOOP_NODE_MAX_COUNT]: process.env.NEXT_PUBLIC_LOOP_NODE_MAX_COUNT,
     [DatasetAttr.DATA_PUBLIC_MAX_ITERATIONS_NUM]: process.env.NEXT_PUBLIC_MAX_ITERATIONS_NUM,
     [DatasetAttr.DATA_PUBLIC_MAX_TREE_DEPTH]: process.env.NEXT_PUBLIC_MAX_TREE_DEPTH,
+    [DatasetAttr.DATA_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME]: process.env.NEXT_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME,
     [DatasetAttr.DATA_PUBLIC_ENABLE_WEBSITE_JINAREADER]: process.env.NEXT_PUBLIC_ENABLE_WEBSITE_JINAREADER,
     [DatasetAttr.DATA_PUBLIC_ENABLE_WEBSITE_FIRECRAWL]: process.env.NEXT_PUBLIC_ENABLE_WEBSITE_FIRECRAWL,
     [DatasetAttr.DATA_PUBLIC_ENABLE_WEBSITE_WATERCRAWL]: process.env.NEXT_PUBLIC_ENABLE_WEBSITE_WATERCRAWL,

web/config/index.ts

@@ -270,6 +270,7 @@ export const LOOP_NODE_MAX_COUNT = getNumberConfig(process.env.NEXT_PUBLIC_LOOP_
 export const MAX_ITERATIONS_NUM = getNumberConfig(process.env.NEXT_PUBLIC_MAX_ITERATIONS_NUM, DatasetAttr.DATA_PUBLIC_MAX_ITERATIONS_NUM, 99)
 export const MAX_TREE_DEPTH = getNumberConfig(process.env.NEXT_PUBLIC_MAX_TREE_DEPTH, DatasetAttr.DATA_PUBLIC_MAX_TREE_DEPTH, 50)
 
+export const ALLOW_UNSAFE_DATA_SCHEME = getBooleanConfig(process.env.NEXT_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME, DatasetAttr.DATA_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME, false)
 export const ENABLE_WEBSITE_JINAREADER = getBooleanConfig(process.env.NEXT_PUBLIC_ENABLE_WEBSITE_JINAREADER, DatasetAttr.DATA_PUBLIC_ENABLE_WEBSITE_JINAREADER, true)
 export const ENABLE_WEBSITE_FIRECRAWL = getBooleanConfig(process.env.NEXT_PUBLIC_ENABLE_WEBSITE_FIRECRAWL, DatasetAttr.DATA_PUBLIC_ENABLE_WEBSITE_FIRECRAWL, true)
 export const ENABLE_WEBSITE_WATERCRAWL = getBooleanConfig(process.env.NEXT_PUBLIC_ENABLE_WEBSITE_WATERCRAWL, DatasetAttr.DATA_PUBLIC_ENABLE_WEBSITE_WATERCRAWL, false)

web/docker/entrypoint.sh

@@ -26,6 +26,7 @@ export NEXT_TELEMETRY_DISABLED=${NEXT_TELEMETRY_DISABLED}
 export NEXT_PUBLIC_TEXT_GENERATION_TIMEOUT_MS=${TEXT_GENERATION_TIMEOUT_MS}
 export NEXT_PUBLIC_CSP_WHITELIST=${CSP_WHITELIST}
 export NEXT_PUBLIC_ALLOW_EMBED=${ALLOW_EMBED}
+export NEXT_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME=${ALLOW_UNSAFE_DATA_SCHEME:-false}
 export NEXT_PUBLIC_TOP_K_MAX_VALUE=${TOP_K_MAX_VALUE}
 export NEXT_PUBLIC_INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH=${INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH}
 export NEXT_PUBLIC_MAX_TOOLS_NUM=${MAX_TOOLS_NUM}

web/types/feature.ts

@@ -116,6 +116,7 @@ export enum DatasetAttr {
   DATA_PUBLIC_LOOP_NODE_MAX_COUNT = 'data-public-loop-node-max-count',
   DATA_PUBLIC_MAX_ITERATIONS_NUM = 'data-public-max-iterations-num',
   DATA_PUBLIC_MAX_TREE_DEPTH = 'data-public-max-tree-depth',
+  DATA_PUBLIC_ALLOW_UNSAFE_DATA_SCHEME = 'data-public-allow-unsafe-data-scheme',
   DATA_PUBLIC_ENABLE_WEBSITE_JINAREADER = 'data-public-enable-website-jinareader',
   DATA_PUBLIC_ENABLE_WEBSITE_FIRECRAWL = 'data-public-enable-website-firecrawl',
   DATA_PUBLIC_ENABLE_WEBSITE_WATERCRAWL = 'data-public-enable-website-watercrawl',