fix(tui-go): make the write/edit output-panel number meaningful

The collapsed tool row showed "tool output (N chars)" where N was len() of the
raw unified diff — i.e. the diff blob's *byte* length (headers, @@, +/-/context
all counted), labelled "chars". That number corresponded to nothing the operator
cares about, and len() is bytes not characters.

- Summarise a write/edit by the diff's row count instead — "· diff (7 rows) —
  ^x to view" — matching exactly what the ^x modal reports. Non-diff output (the
  defensive fallback) now counts runes, not bytes, so "N chars" is truthful.
  Retire itoaLen (the byte-count-as-"len" footgun).
- Harden the (+a −b) line counts in diffSummary: guard the file header with the
  trailing space ("+++ "/"--- "), so a changed line whose content starts with
  "++"/"--" is counted instead of mistaken for a header. Apply the same guard in
  parseUnifiedDiff so such a line is rendered (and counted) rather than dropped.

tool_summary_test.go covers the diff-row summary, the rune (not byte) count, and
the ++/-- content-line edge case.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

apps/tui-go/internal/app/view.go

@@ -589,7 +589,7 @@ func (m Model) buildTranscriptRows(w int) ([]string, []int) {
 				rows = append(rows, t.span(s, t.P.Faint))
 			}
 		case "tool":
-			rows = append(rows, t.span("· tool output ("+itoaLen(e.Content)+" chars) — ^x to view", t.P.Dim))
+			rows = append(rows, t.span(toolRowSummary(e.Content), t.P.Dim))
 		case "narration_llm":
 			diamond := lipgloss.NewStyle().Foreground(t.P.Accent).Background(t.P.Bg).Bold(true).Render("◆")
 			lines := wrap(e.Content, w-2)
@@ -856,7 +856,17 @@ func itoa(n int) string {
 	return string(b[i:])
 }
 
-func itoaLen(s string) string { return itoa(len(s)) }
+// toolRowSummary collapses a tool-output transcript entry into a one-line pointer. A
+// write/edit entry holds a unified diff, so summarise it by what ^x actually shows — the
+// diff's row count — instead of the raw diff byte length (which read as a meaningless
+// "N chars": it counted +/-/@@/header bytes, not anything the operator cares about, and
+// len() is bytes not characters). Non-diff output falls back to a true character count.
+func toolRowSummary(content string) string {
+	if isUnifiedDiff(content) {
+		return "· diff (" + itoa(previewRowCount(content)) + " rows) — ^x to view"
+	}
+	return "· tool output (" + itoa(len([]rune(content))) + " chars) — ^x to view"
+}
 
 // metricsSuffix formats the faint latency+token annotation for a ROUTER turn.
 func metricsSuffix(mtr *TurnMetrics) string {