perf: 3-tier conversation context with LRU cache, keyword relevance, debounced I/O

UPGRADE from naive JSON to production-grade conversation memory:

Tier 1 — Compressed Summary (max 600 chars):
  Incrementally built from evicted messages. Preserves conversation
  topics across 100+ messages in a tiny budget.

Tier 2 — Relevant Snippets (BM25-style keyword matching):
  Scores older messages against current query, injects top 3 matches.
  Zero external deps — keyword extraction is ~0.1ms.

Tier 3 — Sliding Window (last 12 exchanges verbatim):
  Recent context preserved word-for-word, fitting within token budget.

Performance optimizations:
  - In-memory Map cache with lazy-load from disk (0ms reads)
  - Debounced async disk writes (3s, non-blocking, never stalls response)
  - LRU eviction for cache (max 50 chats, prevents memory leak)
  - Keywords stripped before saving (smaller JSON files)
  - Backward-compatible: loads old format without keywords, backfills on load
  - Graceful shutdown flushes all pending saves to disk
  - Token-aware budget allocation: summary 15% + relevant 15% + recent 70%

src/bot/index.js

@@ -688,7 +688,7 @@ export async function initBot(config, api, tools, skills, agents) {
 
       // ── Load conversation history for this chat ──
       const chatKey = conversation._key(ctx.chat.id, ctx.message?.message_thread_id);
-      const history = await conversation.getContext(chatKey);
+      const history = await conversation.getContext(chatKey, text);
 
       // Create stream consumer for real-time edit-in-place
       const consumer = new StreamConsumer(ctx, { editInterval: 1000 });
@@ -757,6 +757,15 @@ export async function initBot(config, api, tools, skills, agents) {
     logger.error('Unhandled rejection:', reason?.message || reason);
   });
 
+  // ── Graceful shutdown: flush conversation history ──
+  const shutdown = async (signal) => {
+    logger.info(`🛑 Shutting down (${signal})...`);
+    await conversation.flush();
+    process.exit(0);
+  };
+  process.on('SIGINT', () => shutdown('SIGINT'));
+  process.on('SIGTERM', () => shutdown('SIGTERM'));
+
   // ── Express + WebSocket server (keep for webhook compatibility) ──
   const app = express();
   app.use(express.json());