Merge branch 'main' into claude/performance-optimization-round-6

Resolves the one conflict in file_blob_read_repository.rs's
suggest_files_by_name: main added the CALLER_CAN_READ_DRIVE authz scope
(caller_id param + drive-membership filter, AuthZ audit finding #1 — the
suggest query previously leaked names/paths across tenants), round 6
switched the same query's id/folder_id columns to binary UUID decode.
Kept both: main's authz structure (format! + CALLER_CAN_READ_DRIVE +
caller_id bind) with round 6's binary decode (fi.id / fi.folder_id, no
::text) so the query matches the FileRow = (Uuid, …) tuple. The
deliberately-text sites (min(fm.file_id::text), folder path lookup)
stay text. Verified: build + clippy -D warnings clean, 524 unit +
554 integration tests green.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017aJu9ghvuT8WqC31ZEGTBA

src/common/stubs.rs

@@ -725,6 +725,7 @@ impl SearchUseCase for StubSearchUseCase {
         _query: &str,
         _folder_id: Option<&str>,
         _limit: usize,
+        _caller_id: Uuid,
     ) -> Result<SearchSuggestionsDto, DomainError> {
         Ok(SearchSuggestionsDto {
             suggestions: Vec::new(),