Add Phase 6: write tools, shell, and edit tools with reasoning-only fix

Implement 6 new agent tools — write_file, make_dir, delete_file,
str_replace, patch_apply, run_command — bringing the agent from
read-only observer to active code modifier. All write/shell operations
are gated through the existing permissions service.

Also fix a bug where qwen3.5 thinking mode produces reasoning tokens
but no content after tool results, causing the agent to silently exit.
The loop now detects reasoning-only responses, retries twice, then
injects a nudge message to break the model out of its thinking loop.

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

app/services/streaming.py

@@ -147,6 +147,11 @@ class StreamHandler:
         """Token usage reported by the API, if available."""
         return self._usage
 
+    @property
+    def had_reasoning_only(self) -> bool:
+        """True if the model produced reasoning tokens but no content or tool calls."""
+        return bool(self._accumulated_reasoning) and not self._accumulated_content and not self._tool_calls
+
     def reset(self) -> None:
         """Clear all accumulators for the next turn."""
         self._accumulated_content = ""