feat(llm): improve UTF-8 handling in response processing to prevent infinite loops

feat(shell): enhance output capturing by using threads for stdout and stderr
feat(tui): update usage widget to display token counts and provider/model information
refactor(tui): simplify status bar rendering by removing unnecessary token calculations

apps/infrastructure/src/llm/provider.rs

@@ -386,13 +386,22 @@ impl LlmClient {
                 break;
             }
             byte_buf.extend_from_slice(&chunk_buf[..n]);
+
+            // Drain any bytes that are not valid UTF-8 to prevent
+            // infinite loop when a non-UTF-8 sequence is received.
             let valid_len = match std::str::from_utf8(&byte_buf) {
                 Ok(s) => s.len(),
-                Err(e) => e.valid_up_to(),
+                Err(e) => {
+                    let n = e.valid_up_to();
+                    if n == 0 {
+                        // No valid UTF-8 prefix; skip the first byte (likely
+                        // a partial multi-byte sequence or stray byte).
+                        byte_buf.drain(..1);
+                        continue;
+                    }
+                    n
+                }
             };
-            if valid_len == 0 {
-                continue;
-            }
             let text =
                 String::from_utf8_lossy(&byte_buf[..valid_len]).into_owned();
             byte_buf.drain(..valid_len);

apps/infrastructure/src/tools/shell.rs

@@ -89,30 +89,47 @@ impl Tool for Bash {
         let start = std::time::Instant::now();
         let timeout = Duration::from_millis(timeout_ms);
 
+        let mut child_stdout = child.stdout.take()
+            .ok_or_else(|| anyhow::anyhow!("failed to capture stdout"))?;
+        let mut child_stderr = child.stderr.take()
+            .ok_or_else(|| anyhow::anyhow!("failed to capture stderr"))?;
+
+        let stdout_handle = std::thread::spawn(move || -> std::io::Result<Vec<u8>> {
+            use std::io::Read;
+            let mut buf = Vec::new();
+            child_stdout.read_to_end(&mut buf)?;
+            Ok(buf)
+        });
+        let stderr_handle = std::thread::spawn(move || -> std::io::Result<Vec<u8>> {
+            use std::io::Read;
+            let mut buf = Vec::new();
+            child_stderr.read_to_end(&mut buf)?;
+            Ok(buf)
+        });
+
         loop {
             match child.try_wait() {
                 Ok(Some(status)) => {
+                    let stdout = stdout_handle.join().unwrap_or(Ok(Vec::new()))?;
+                    let stderr = stderr_handle.join().unwrap_or(Ok(Vec::new()))?;
+
                     let elapsed = start.elapsed().as_secs_f64();
-                    let output = child
+                    let stdout_str = String::from_utf8_lossy(&stdout).to_string();
-                        .wait_with_output()
+                    let stderr_str = String::from_utf8_lossy(&stderr).to_string();
-                        .map_err(|e| anyhow::anyhow!("failed to collect output: {e}"))?;
+                    let combined = if stderr_str.is_empty() {
-                    let stdout = String::from_utf8_lossy(&output.stdout).to_string();
+                        stdout_str
-                    let stderr = String::from_utf8_lossy(&output.stderr).to_string();
-                    let combined = if stderr.is_empty() {
-                        stdout
                     } else {
-                        format!("{stdout}\n{stderr}")
+                        format!("{stdout_str}\n{stderr_str}")
                     };
                     let trimmed = combined.trim().to_string();
+
                     if status.success() {
-                        debug!(elapsed_secs = elapsed, "bash command completed successfully");
                         return Ok(if trimmed.is_empty() {
                             format!("Command completed in {elapsed:.2}s (exit code 0)")
                         } else {
                             format!("{trimmed}\n\nExit code: 0 ({elapsed:.2}s)")
                         });
                     }
-                    warn!(exit_code = status.code().unwrap_or(-1), elapsed_secs = elapsed, "bash command failed");
                     return Ok(format!(
                         "{}\n\nExit code: {} ({:.2}s)",
                         trimmed,
@@ -124,7 +141,6 @@ impl Tool for Bash {
                     if start.elapsed() > timeout {
                         let _ = child.kill();
                         let _ = child.wait();
-                        warn!(timeout_ms, "bash command timed out");
                         anyhow::bail!("command timed out after {timeout_ms}ms");
                     }
                     std::thread::sleep(Duration::from_millis(10));

apps/interfaces/tui/src/view/sidebar.rs

@@ -96,19 +96,40 @@ fn draw_usage_widget(frame: &mut Frame, area: Rect, state: &crate::state::AppSta
     let lines: Vec<Line> = if let Some(ref rt) = state.session_runtime {
         let now_ms = chrono::Utc::now().timestamp_millis();
         let summary = compute_usage_summary(&rt.usage, rt.session_start, now_ms);
-        vec![
+        let max_tokens = crate::state::resolve_context_window(&state.app_config, &state.settings);
+        let current_tokens = state.cached_token_count;
+        
+        let mut items = vec![
             Line::from(Span::styled(
                 format!("  {:>6}: {} tok", "total", summary.total_tokens),
                 Style::default()
                     .fg(Theme::TEXT)
                     .add_modifier(Modifier::BOLD),
             )),
-            Line::from(Span::styled(
+        ];
+
+        if summary.self_learning_tokens > 0 {
+            items.push(Line::from(Span::styled(
                 format!("  {:>6}: {} tok", "main", summary.main_tokens),
                 Style::default().fg(Theme::TEXT_DIM),
+            )));
+            items.push(Line::from(Span::styled(
+                format!("  {:>6}: {} tok", "learn", summary.self_learning_tokens),
+                Style::default().fg(Theme::TEXT_DIM),
+            )));
+        }
+
+        items.extend(vec![
+            Line::from(Span::styled(
+                format!("  {:>6}: {}/{}", "ctx", current_tokens, max_tokens),
+                Style::default().fg(Theme::TEXT_DIM),
             )),
             Line::from(Span::styled(
-                format!("  {:>6}: {} tok", "learn", summary.self_learning_tokens),
+                format!("  {:>6}: {}", "prov", state.settings.provider),
+                Style::default().fg(Theme::TEXT_DIM),
+            )),
+            Line::from(Span::styled(
+                format!("  {:>6}: {}", "model", state.settings.model),
                 Style::default().fg(Theme::TEXT_DIM),
             )),
             Line::from(Span::styled(
@@ -122,7 +143,9 @@ fn draw_usage_widget(frame: &mut Frame, area: Rect, state: &crate::state::AppSta
                 ),
                 Style::default().fg(Theme::TEXT_DIM),
             )),
-        ]
+        ]);
+        
+        items
     } else {
         vec![Line::from(Span::styled(
             " No active session.",

apps/interfaces/tui/src/view/status.rs

@@ -48,32 +48,7 @@ pub fn draw_status_bar(frame: &mut Frame, area: Rect, state: &crate::state::AppS
         status_badge,
     ];
 
-    let max_tokens = crate::state::resolve_context_window(&state.app_config, &state.settings);
+    let right_str = String::new();
-
-    let right_str = if let Some(ref rt) = state.session_runtime {
-        // Use cached token count — recomputed lazily only when new messages
-        // arrive (token_count_dirty flag), not on every render frame.
-        // This eliminates the expensive tiktoken_rs call from the hot path.
-        let current_tokens = state.cached_token_count;
-
-        let mut parts = Vec::new();
-        if rt.usage.last_tokens_in > 0 || rt.usage.last_tokens_out > 0 {
-            parts.push(format!(
-                "↑{} ↓{}",
-                rt.usage.last_tokens_in, rt.usage.last_tokens_out
-            ));
-        }
-        parts.push(format!("{current_tokens}/{max_tokens}"));
-        parts.push(state.settings.provider.clone());
-        parts.push(state.settings.model.clone());
-
-        format!(" {} ", parts.join(" · "))
-    } else {
-        format!(
-            " 0/{max_tokens} · {} · {} ",
-            state.settings.provider, state.settings.model
-        )
-    };
 
     let left_line = Line::from(left_spans);
     let right_line = Line::from(Span::styled(