fix: Eliminate blocking HTTP call from instances_use

instances_use previously called register_instance which made a
blocking safe_get (30s timeout) to validate the connection. If the
Ghidra server was slow or unresponsive, this could hang the MCP tool
call indefinitely from the client's perspective.

Now instances_use creates a lazy stub entry and sets the port
immediately — pure in-memory, no network I/O. The first actual tool
call validates the connection naturally.

Also fix background discovery thread using request_timeout (30s)
instead of discovery_timeout (0.5s) per port — worst case went from
300s to 5s per scan cycle.

src/ghydramcp/mixins/instances.py

@@ -166,13 +166,20 @@ class InstancesMixin(GhydraMixinBase):
         Returns:
             Confirmation message with instance details
         """
+        # Register lazily without blocking HTTP calls.
+        # If the instance is unknown, create a stub entry — the first
+        # actual tool call (functions_list, etc.) will validate the
+        # connection and fail fast with a clear error if unreachable.
         with self._instances_lock:
-            needs_register = port not in self._instances
-
-        if needs_register:
-            result = self.register_instance(port)
-            if "Failed" in result or "Error" in result:
-                return result
+            if port not in self._instances:
+                config = get_config()
+                self._instances[port] = {
+                    "url": f"http://{config.ghidra_host}:{port}",
+                    "project": "",
+                    "file": "",
+                    "registered_at": time.time(),
+                    "lazy": True,
+                }
 
         self.set_current_port(port)
 

src/ghydramcp/server.py

@@ -105,10 +105,14 @@ def create_server(
 def _periodic_discovery(interval: int = 30):
     """Background thread for periodic instance discovery.
 
+    Uses a short timeout per port so a full scan completes quickly
+    even when most ports are unreachable.
+
     Args:
         interval: Seconds between discovery attempts
     """
-    from .core.http_client import safe_get
+    import requests as _requests
+
     from .mixins.base import GhydraMixinBase
 
     config = get_config()
@@ -116,15 +120,22 @@ def _periodic_discovery(interval: int = 30):
     while True:
         time.sleep(interval)
         try:
-            # Quick scan of common ports
+            # Quick scan — use discovery_timeout (0.5s), NOT request_timeout (30s)
             for port in config.quick_discovery_range:
                 try:
-                    response = safe_get(port, "")
+                    url = f"http://{config.ghidra_host}:{port}/"
+                    resp = _requests.get(
+                        url,
+                        timeout=config.discovery_timeout,
+                        headers={"Accept": "application/json"},
+                    )
+                    if resp.ok:
+                        response = resp.json()
                         if response.get("success", False):
                             with GhydraMixinBase._instances_lock:
                                 if port not in GhydraMixinBase._instances:
                                     GhydraMixinBase._instances[port] = {
-                                    "url": f"http://{config.ghidra_host}:{port}",
+                                        "url": url.rstrip("/"),
                                         "project": response.get("project", ""),
                                         "file": response.get("file", ""),
                                         "discovered_at": time.time(),