fix: Disable spinner animation to stay within rate limits (v1.6.4)

CRITICAL FIX: Animated spinner was causing excessive API calls.

## Problem

Each editMessageText() counts as 1 API call toward Telegram's
20 msgs/min per-chat limit. The spinner animation updated every
1 second, which means:

- 5 tools × 10 seconds = 50 seconds execution
- Spinner updates: 50 calls
- Tool status changes: 10 calls
- Text updates: ~5 calls
- Total: ~65 calls/min → EXCEEDS 20/min limit!

Even at 3-second intervals:
- Spinner: ~16 calls
- Other: ~15 calls
- Total: ~31 calls/min → Still exceeds limit!

## Solution

**Disable spinner animation entirely.** The tool overview now only
updates when tool status actually changes (start/done), not on a
timer.

Updates per minute:
- Tool status changes: ~10 calls
- Text updates: ~5 calls
- Done message: 1 call
- Total: ~16 calls/min ✅ Within 18/min limit

## Impact

**Before (with spinner):**
- Tool overview updates ~50 times/min
- Risk of hitting rate limit
- Could cause 429 errors

**After (no spinner):**
- Tool overview updates only on status change (~10 times)
- Well within rate limits
- No 429 risk

## User Experience

Still responsive because:
- Tool status changes update IMMEDIATELY
- Counter updates in real-time (0/4 → 1/4 → 2/4)
- No perceived delay since updates happen on meaningful events
- Users see progress, just not cosmetic spinner animation

## Key Insight

Telegram counts EVERY API call:
- sendMessage = 1 call
- editMessageText = 1 call ⭐️
- deleteMessage = 1 call

There's no "free" edit operation. Even simple cosmetic updates
consume the rate limit budget.

## Files Changed
- src/telegram-message-queue.ts: Disable spinner, update only on status change
- doc/rate-limiting.md: Document why spinner is disabled

Prevents rate limit exhaustion from cosmetic updates.

Author: htlin222

doc/rate-limiting.md

@@ -60,14 +60,14 @@ Intelligently batches and merges messages by priority:
 
 **After:**
 ```
-⚙️ Tools
+🔧 Tools
 🔧 Reading file.ts...
 📝 Editing config.json...
 🔍 Grepping pattern...
 ✓ Writing output.txt
 2/4 complete
 ```
-1 merged message with animated spinner (⚙️ → 🔧 → ⚡ → 💫), updated immediately when status changes
+1 merged message, updated immediately when tool status changes (no spinner to avoid rate limits)
 
 ### 3. Integration in `streaming.ts`
 
@@ -218,13 +218,15 @@ To test with heavy tool usage:
 
 ## Responsiveness Optimizations
 
-To prevent "stuck" feeling:
+To prevent "stuck" feeling while staying within rate limits:
 
 1. **Tool updates are immediate** - Not batched, updated as soon as status changes
-2. **Animated spinner** - Cycles through emoji (⚙️ → 🔧 → ⚡ → 💫) every 1s to show activity
+2. **No animated spinner** - Removed to avoid excessive API calls (each edit = 1 API call)
 3. **Shorter delays** - Text batching uses 1.5s instead of 3s
 4. **Smart rate limiting** - Only delays when actually approaching limits
 
+**Why no spinner?** Each `editMessageText()` counts as 1 API call toward the 20 msgs/min limit. Updating a spinner every 1-3 seconds would quickly exhaust the limit. Instead, the overview updates immediately when tool status changes (start/done), providing real-time feedback without excessive API calls.
+
 ## Future Improvements
 
 Potential enhancements:

package.json

@@ -1,6 +1,6 @@
 {
 	"name": "ctb",
-	"version": "1.6.3",
+	"version": "1.6.4",
 	"description": "Control Claude Code from Telegram - run multiple bot instances per project",
 	"type": "module",
 	"bin": {

src/telegram-message-queue.ts

@@ -294,34 +294,20 @@ export class TelegramMessageQueue {
 
 	/**
 	 * Start spinner animation for tool overview.
+	 *
+	 * IMPORTANT: Spinner updates are DISABLED to avoid rate limiting.
+	 * Each editMessageText counts as an API call, and updating every few
+	 * seconds would quickly exhaust the 18 msgs/min per-chat limit.
+	 *
+	 * Instead, we only update when tool status actually changes.
 	 */
 	private startToolSpinner(ctx: Context, chatId: number): void {
 		// Stop existing spinner if any
 		this.stopToolSpinner(chatId);
 
-		const spinnerFrames = ["⚙️", "🔧", "⚡", "💫"];
-		let frameIndex = 0;
-
-		const interval = setInterval(async () => {
-			const tools = this.activeTools.get(chatId);
-			if (!tools || tools.filter((t) => t.status === "running").length === 0) {
-				this.stopToolSpinner(chatId);
-				return;
-			}
-
-			frameIndex = (frameIndex + 1) % spinnerFrames.length;
-			await this.updateToolOverviewMessage(
-				ctx,
-				chatId,
-				tools,
-				spinnerFrames[frameIndex],
-			).catch(() => {
-				// Stop on error
-				this.stopToolSpinner(chatId);
-			});
-		}, 1000); // Update spinner every 1 second
-
-		this.toolSpinners.set(chatId, interval);
+		// Spinner animation is DISABLED - see comment above
+		// We rely on tool status changes to update the overview message
+		// This keeps us well within rate limits
 	}
 
 	/**