fix: daemon delete creates tombstones, export includes tombstones (bd-rp4o)

Three changes to fix deleted issues resurrecting during bd sync:

1. daemon handleDelete now uses CreateTombstone instead of DeleteIssue
   - internal/rpc/server_issues_epics.go

2. sync.go exportToJSONL now includes IncludeTombstones:true
   - cmd/bd/sync.go

3. server_export_import_auto.go handleExport and auto-export now include
   tombstones in SearchIssues filter
   - internal/rpc/server_export_import_auto.go

Also adds README.md documentation for sync.branch mode (bd-dsdh):
- Explains "always dirty" working tree behavior
- Shell alias tip: gs='git status -- ":!.beads/"'
- When to use bd sync --merge

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

internal/rpc/server_export_import_auto.go

@@ -49,8 +49,9 @@ func (s *Server) handleExport(req *Request) Response {
 		manifest = export.NewManifest(cfg.Policy)
 	}
 
-	// Get all issues (core operation, always fail-fast)
-	issues, err := store.SearchIssues(ctx, "", types.IssueFilter{})
+	// Get all issues including tombstones for sync propagation (bd-rp4o fix)
+	// Tombstones must be exported so they propagate to other clones and prevent resurrection
+	issues, err := store.SearchIssues(ctx, "", types.IssueFilter{IncludeTombstones: true})
 	if err != nil {
 		return Response{
 			Success: false,
@@ -464,8 +465,8 @@ func (s *Server) triggerExport(ctx context.Context, store storage.Storage, dbPat
 		}
 	}
 
-	// Export to JSONL (this will update the file with remapped IDs)
-	allIssues, err := sqliteStore.SearchIssues(ctx, "", types.IssueFilter{})
+	// Export to JSONL including tombstones for sync propagation (bd-rp4o fix)
+	allIssues, err := sqliteStore.SearchIssues(ctx, "", types.IssueFilter{IncludeTombstones: true})
 	if err != nil {
 		return fmt.Errorf("failed to fetch issues for export: %w", err)
 	}