fix(knowledge): re-index a file when an ancestor folder is renamed

Cursor Bugbot, medium severity. The hash already carried folderId so a move
re-indexes, but the resolved folderPath is what gets stored as the tag, and
renaming an ancestor folder rewrites that path while writing only the folder
table. contentUpdatedAt, originalName and folderId all stay put, so
classifyExternalDoc saw no change and the document kept the old folder tag
indefinitely.

folderPath now participates in the hash. Both listDocuments and getDocument
resolve it from the same pathById map, so the listing and hydration phases
still produce byte-identical hashes. Mutation-checked: dropping it from the
hash fails the new test.

Author: mzxchandra

apps/sim/connectors/sim-files/sim-files.test.ts

@@ -127,6 +127,17 @@ describe('fileRowToStub', () => {
     expect(after).toBe(before)
   })
 
+  /**
+   * Renaming an ancestor folder rewrites the stored path tag but writes only the
+   * `folder` table — `contentUpdatedAt`, `originalName` and `folderId` all stay put,
+   * so without the path in the hash the document keeps the old folder tag forever.
+   */
+  it('changes the hash when an ancestor folder rename moves the path', () => {
+    expect(fileRowToStub(BASE_ROW, 'ws-1', 'Documentation/Specs').contentHash).not.toBe(
+      fileRowToStub(BASE_ROW, 'ws-1', 'Docs/Specs').contentHash
+    )
+  })
+
   /** listDocuments and getDocument must produce byte-identical hashes for one row. */
   it('is deterministic for the same row', () => {
     expect(fileRowToStub(BASE_ROW, 'ws-1', 'Docs/Specs').contentHash).toBe(

apps/sim/connectors/sim-files/sim-files.ts

@@ -186,6 +186,12 @@ export function buildFileListingFilters(args: {
  * `originalName` and `folderId` participate because a rename or a move alters what
  * we index (title, folder tag) without touching `contentUpdatedAt`, which advances
  * only on content writes.
+ *
+ * `folderPath` participates for the same reason one level up: renaming an ANCESTOR
+ * folder rewrites the path we store as a tag but writes only the `folder` table, so
+ * every other hashed field stays put and the document would keep the old path
+ * forever. Both `listDocuments` and `getDocument` resolve it from the same
+ * `pathById` map, so the two phases still agree.
  */
 export function fileRowToStub(
   row: FileRow,
@@ -201,7 +207,7 @@ export function fileRowToStub(
     sourceUrl: `${getBaseUrl()}/workspace/${workspaceId}/files${
       row.folderId ? `?folderId=${encodeURIComponent(row.folderId)}` : ''
     }`,
-    contentHash: `simfile:${row.id}:${row.contentUpdatedAt.toISOString()}:${row.originalName}:${row.folderId ?? ''}`,
+    contentHash: `simfile:${row.id}:${row.contentUpdatedAt.toISOString()}:${row.originalName}:${row.folderId ?? ''}:${folderPath ?? ''}`,
     metadata: {
       folderPath: folderPath ?? '',
       contentType: row.contentType,