Fix doctor incorrectly diagnosing hash IDs as sequential (issue #322)

- Enhanced checkIDFormat to sample multiple issues instead of just one
- Added detectHashBasedIDs function with robust multi-heuristic detection:
  * Checks for child_counters table (hash ID schema indicator)
  * Detects letters in IDs (base36 encoding)
  * Identifies leading zeros (common in hash IDs, rare in sequential)
  * Analyzes variable length patterns (adaptive hash IDs)
  * Checks for non-sequential numeric ordering
- Added comprehensive test coverage (16 new test cases)
- Fixes false positives for numeric-only hash IDs like 'pf-0088'

Closes #322

cmd/bd/migrate_hash_ids_test.go

@@ -251,13 +251,33 @@ func TestIsHashID(t *testing.T) {
 		id       string
 		expected bool
 	}{
+		// Sequential IDs (numeric only, short)
 		{"bd-1", false},
 		{"bd-123", false},
+		{"bd-9999", false},
+		
+		// Hash IDs with letters
 		{"bd-a3f8e9a2", true},
 		{"bd-abc123", true},
 		{"bd-123abc", true},
 		{"bd-a3f8e9a2.1", true},
 		{"bd-a3f8e9a2.1.2", true},
+		
+		// Hash IDs that are numeric but 5+ characters (likely hash)
+		{"bd-12345", true},
+		{"bd-0088", false}, // 4 chars, all numeric - ambiguous, defaults to false
+		{"bd-00880", true},  // 5+ chars, likely hash
+		
+		// Base36 hash IDs with letters
+		{"bd-5n3", true},
+		{"bd-65w", true},
+		{"bd-jmx", true},
+		{"bd-4rt", true},
+		
+		// Edge cases
+		{"bd-", false},     // Empty suffix
+		{"invalid", false}, // No dash
+		{"bd-0", false},    // Single digit
 	}
 
 	for _, tt := range tests {