docs: fix inaccuracies found during documentation accuracy review

- CHANGELOG: inactivity_kill_secs default is disabled (None), not 120
- TROUBLESHOOTING: clarify inactivity_kill_secs is opt-in (must be
  passed explicitly); correct "default 120" claim
- MCP_TOOLS: inactivity_kill_secs is disabled by default; add note
  that Wine/macOS may not exit naturally even with ShutdownTerminal=1
- handlers/backtest.rs: remove stale comment claiming inactivity
  watchdog is "intentionally skipped when shutdown=true" — replaced
  with accurate description of the new 30s HTML-wait + kill behavior

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/tools/handlers/backtest.rs

@@ -338,10 +338,11 @@ pub async fn handle_launch_backtest(handler: &crate::tools::handlers::ToolHandle
         skip_clean: args.get("skip_clean").and_then(|v| v.as_bool()).unwrap_or(false),
         skip_analyze: true, // Not needed for launch mode
         deep_analyze: false,
-        // ShutdownTerminal=1 (default): MT5 writes the HTML report and exits cleanly.
-        // The background monitor's post-exit scan finds the report within 10s.
-        // The inactivity watchdog is intentionally skipped when shutdown=true so it
-        // doesn't race with MT5's report write right before natural exit.
+        // On Wine/macOS, ShutdownTerminal=1 does NOT reliably cause terminal64.exe to exit.
+        // When inactivity_kill_secs is set, the monitor waits that many seconds after the
+        // tester log goes quiet, then polls for the HTML report for 30 s, then kills MT5.
+        // If no inactivity_kill_secs is given (default None → disabled), the monitor relies
+        // solely on timeout (900 s) or natural MT5 exit for completion detection.
         shutdown: args.get("shutdown").and_then(|v| v.as_bool()).unwrap_or(true),
         kill_existing: false,
         timeout: args.get("timeout").and_then(|v| v.as_u64()).unwrap_or(900),