fix(files): normalize encoded embedded ids (#7035)

Author: waleedlatif1

apps/sim/app/api/files/export/[id]/route.test.ts

@@ -33,7 +33,6 @@ vi.mock('@/app/api/files/authorization', () => ({ verifyFileAccess: mockVerifyFi
 vi.mock('@/lib/uploads/core/storage-service', () => ({ downloadFile: mockDownloadFile }))
 vi.mock('@/lib/uploads/server/embedded-image-refs', () => ({
   extractEmbeddedFileRefs: mockExtractEmbeddedFileRefs,
-  storedFileId: (spelledId: string) => decodeURIComponent(spelledId),
 }))
 vi.mock('@sim/audit', () => ({
   recordAudit: vi.fn(),

apps/sim/app/api/files/export/[id]/route.ts

@@ -15,8 +15,9 @@ import { captureServerEvent } from '@/lib/posthog/server'
 import type { StorageContext } from '@/lib/uploads/config'
 import { getServeStoragePrefix } from '@/lib/uploads/config'
 import { downloadFile } from '@/lib/uploads/core/storage-service'
-import { extractEmbeddedFileRefs, storedFileId } from '@/lib/uploads/server/embedded-image-refs'
+import { extractEmbeddedFileRefs } from '@/lib/uploads/server/embedded-image-refs'
 import { getFileMetadataById } from '@/lib/uploads/server/metadata'
+import { storedFileId } from '@/lib/uploads/utils/embedded-image-ref'
 import { formatFileSize } from '@/lib/uploads/utils/file-utils'
 import { verifyFileAccess } from '@/app/api/files/authorization'
 import { encodeFilenameForHeader } from '@/app/api/files/utils'

apps/sim/app/api/files/public/[token]/inline/route.test.ts

@@ -67,6 +67,15 @@ describe('GET /api/files/public/[token]/inline', () => {
     expect(res.headers.get('content-type')).toBe('image/png')
   })
 
+  it('serves an image whose id is percent-encoded in the document', async () => {
+    mockDownloadFile.mockImplementation(downloadByKey('![a](/api/files/view/wf%5Fabc)'))
+
+    const res = await GET(req('fileId=wf%5Fabc'), params)
+
+    expect(res.status).toBe(200)
+    expect(mockResolveImage).toHaveBeenCalledWith('ws-1', { fileId: 'wf_abc' })
+  })
+
   it('serves a key-referenced image', async () => {
     mockDownloadFile.mockImplementation(
       downloadByKey(`![a](/api/files/serve/${encodeURIComponent(IMG_KEY)}?context=workspace)`)

apps/sim/app/api/files/public/[token]/inline/route.ts

@@ -12,6 +12,7 @@ import { resolveActiveShareByToken } from '@/lib/public-shares/share-manager'
 import { downloadFile } from '@/lib/uploads/core/storage-service'
 import { extractEmbeddedFileRefs } from '@/lib/uploads/server/embedded-image-refs'
 import { resolveWorkspaceInlineImage } from '@/lib/uploads/server/inline-image'
+import { storedFileId } from '@/lib/uploads/utils/embedded-image-ref'
 import { serveInlineImage } from '@/app/api/files/serve-inline-image'
 import { createErrorResponse, FileNotFoundError } from '@/app/api/files/utils'
 
@@ -73,7 +74,9 @@ export const GET = withRouteHandler(
       // Referenced-by-doc gate: the share grants exactly the images the document embeds.
       const docText = (await downloadFile({ key: doc.key, context: 'workspace' })).toString('utf-8')
       const { keys, ids } = extractEmbeddedFileRefs(docText)
-      const referenced = ref.fileId ? ids.includes(ref.fileId) : keys.includes(ref.key as string)
+      const referenced = ref.fileId
+        ? ids.some((id) => storedFileId(id) === ref.fileId)
+        : keys.includes(ref.key as string)
       if (!referenced) {
         throw new FileNotFoundError('Not found')
       }

apps/sim/app/workspace/[workspaceId]/files/components/file-viewer/rich-markdown-editor/image.test.ts

@@ -28,6 +28,18 @@ describe('content-source resolveImageSrc', () => {
     )
   })
 
+  it('normalizes a percent-encoded id before building an inline request', () => {
+    const workspace = createWorkspaceFileContentSource('ws-1')
+    const publicShare = createPublicFileContentSource('tok_1', '/api/files/public/tok_1/content')
+
+    expect(workspace.resolveImageSrc('/api/files/view/wf%5Fabc')).toBe(
+      '/api/workspaces/ws-1/files/inline?fileId=wf_abc'
+    )
+    expect(publicShare.resolveImageSrc('/api/files/view/wf%5Fabc')).toBe(
+      '/api/files/public/tok_1/inline?fileId=wf_abc'
+    )
+  })
+
   it('passes external/data srcs through unchanged in both sources', () => {
     const ws = createWorkspaceFileContentSource('ws-1')
     const pub = createPublicFileContentSource('tok_1', '/c')

apps/sim/hooks/use-file-content-source.tsx

@@ -4,6 +4,7 @@ import { createContext, useContext } from 'react'
 import {
   type EmbeddedFileRef,
   extractEmbeddedFileRef,
+  storedFileId,
 } from '@/lib/uploads/utils/embedded-image-ref'
 
 export interface FileContentUrlOptions {
@@ -24,7 +25,7 @@ export interface FileContentUrlOptions {
 function inlineRefQuery(ref: NonNullable<EmbeddedFileRef>): string {
   return 'key' in ref
     ? `key=${encodeURIComponent(ref.key)}`
-    : `fileId=${encodeURIComponent(ref.fileId)}`
+    : `fileId=${encodeURIComponent(storedFileId(ref.fileId))}`
 }
 
 export interface ImageDimensions {

apps/sim/lib/copilot/tools/server/files/embedded-image-refs.ts

@@ -1,5 +1,6 @@
-import { extractEmbeddedFileRefs, storedFileId } from '@/lib/uploads/server/embedded-image-refs'
+import { extractEmbeddedFileRefs } from '@/lib/uploads/server/embedded-image-refs'
 import { getFileMetadataById } from '@/lib/uploads/server/metadata'
+import { storedFileId } from '@/lib/uploads/utils/embedded-image-ref'
 
 /**
  * Returns the ids of image embeds in `content` that will not render or survive a workspace export.

apps/sim/lib/uploads/server/embedded-image-refs.ts

@@ -10,24 +10,6 @@ import { extractEmbeddedFileRef, extractImgSrcs } from '@/lib/uploads/utils/embe
 /** Hard cap on embedded images resolved from one document — bounds export bundles and the share cascade. */
 export const MAX_EMBEDDED_IMAGES = 50
 
-/**
- * The stored id behind the spelling a document used. {@link extractEmbeddedFileRefs} returns ids
- * exactly as the document writes them, because that is what the export bundler searches for when it
- * rewrites an embed — but storage is keyed by the decoded id. Every consumer that resolves one of
- * these ids goes through here, so a document's spelling and the stored id stay distinct without
- * either side re-deriving the other.
- *
- * Only for ids read out of document text. Ids arriving as request input are already constrained to
- * the plain id charset by their route contract, so they need no decoding.
- */
-export function storedFileId(spelledId: string): string {
-  try {
-    return decodeURIComponent(spelledId)
-  } catch {
-    return spelledId
-  }
-}
-
 /**
  * A parser of this module's own, not the `marked` singleton: the public share's referenced-by-doc
  * gate authorizes against what this returns, and a global `marked.use()` elsewhere in the process

apps/sim/lib/uploads/utils/embedded-image-ref.test.ts

@@ -1,5 +1,9 @@
 import { describe, expect, it } from 'vitest'
-import { extractEmbeddedFileRef, extractImgSrcs } from '@/lib/uploads/utils/embedded-image-ref'
+import {
+  extractEmbeddedFileRef,
+  extractImgSrcs,
+  storedFileId,
+} from '@/lib/uploads/utils/embedded-image-ref'
 
 const KEY = 'workspace/W1/1700000000000-deadbeefdeadbeef-photo.png'
 const ENCODED = encodeURIComponent(KEY)
@@ -37,6 +41,17 @@ describe('extractEmbeddedFileRef', () => {
   })
 })
 
+describe('storedFileId', () => {
+  it('decodes a document-spelled id exactly once', () => {
+    expect(storedFileId('wf%5Fabc')).toBe('wf_abc')
+    expect(storedFileId('wf%255Fabc')).toBe('wf%5Fabc')
+  })
+
+  it('leaves malformed encodings unchanged', () => {
+    expect(storedFileId('wf%5')).toBe('wf%5')
+  })
+})
+
 describe('extractImgSrcs', () => {
   it('reads double-quoted, single-quoted, and unquoted srcs in document order', () => {
     expect(

apps/sim/lib/uploads/utils/embedded-image-ref.ts

@@ -12,6 +12,19 @@
 /** A reference parsed from an embed `src`: a workspace storage key, a workspace file id, or neither. */
 export type EmbeddedFileRef = { key: string } | { fileId: string } | null
 
+/**
+ * The stored id behind the spelling a document used. Embedded refs retain ids exactly as written so
+ * export rewriting can find the original URL, while storage and inline routes use the decoded id.
+ * Decode exactly once so malformed or double-encoded spellings continue to fail closed.
+ */
+export function storedFileId(spelledId: string): string {
+  try {
+    return decodeURIComponent(spelledId)
+  } catch {
+    return spelledId
+  }
+}
+
 /**
  * Parse a single embed `src` into the workspace file it references, normalizing the spellings the
  * editor and file agent produce: `/api/files/serve/<key>` (incl. `s3/`/`blob/`/`gcs/` prefixes), `/api/files/view/<id>`,