Add passive mesh context awareness — observe channel traffic, inject into LLM prompts

New context.py module: ring buffer (50K hard cap, ~25MB ceiling) passively
records all channel broadcasts. Observations are formatted with relative
timestamps and injected into the system prompt when generating LLM responses.
Only public channel traffic is observed; DMs to the bot are excluded (already
in per-user history). Bot's own node ID is auto-added to ignore list.

Config: context.enabled, observe_channels, ignore_nodes, max_age, max_context_items
TUI: new Context settings submenu (menu item 7)
Hourly prune removes expired observations.

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

meshai/router.py

@@ -10,6 +10,7 @@ from .backends.base import LLMBackend
 from .commands import CommandContext, CommandDispatcher
 from .config import Config
 from .connector import MeshConnector, MeshMessage
+from .context import MeshContext
 from .history import ConversationHistory
 
 logger = logging.getLogger(__name__)
@@ -61,12 +62,14 @@ class MessageRouter:
         history: ConversationHistory,
         dispatcher: CommandDispatcher,
         llm_backend: LLMBackend,
+        context: MeshContext = None,
     ):
         self.config = config
         self.connector = connector
         self.history = history
         self.dispatcher = dispatcher
         self.llm = llm_backend
+        self.context = context
 
 
     def should_respond(self, message: MeshMessage) -> bool:
@@ -147,6 +150,16 @@ class MessageRouter:
         if getattr(self.config.llm, 'use_system_prompt', True):
             system_prompt = self.config.llm.system_prompt
 
+        # Inject mesh context if available
+        if self.context:
+            max_items = getattr(self.config.context, 'max_context_items', 20)
+            context_block = self.context.get_context_block(max_items=max_items)
+            if context_block:
+                system_prompt += (
+                    "\n\n--- Recent mesh traffic (for context only, not messages to you) ---\n"
+                    + context_block
+                )
+
         try:
             response = await self.llm.generate(
                 messages=history,