fix(context): keep tool-output retrieval reliable and stream pages

Exempt read_tool_output from result-bounding so paging a large stored
output isn't re-spilled under a new id, byte-cap each page in place, and
stream the requested window with islice instead of loading the whole
file per page.

tests/test_agent_factory_shell.py

@@ -110,3 +110,14 @@ async def test_chat_completions_filesystem_custom_tool_becomes_function_tool() -
     factory._configure_filesystem_tools(toolset, chat_completions=True)
 
     assert isinstance(toolset.read_file, FunctionTool)
+
+
+def test_read_tool_output_is_not_result_bounded() -> None:
+    # Every other FunctionTool gets the result-bounding wrapper, but the
+    # retrieval tool must be exempt or paging a large stored output would be
+    # re-spilled under a new id.
+    agent = factory.build_strix_agent(is_root=True)
+    by_name = {t.name: t for t in agent.tools}
+
+    assert getattr(by_name["read_tool_output"], "_strix_bounded", False) is False
+    assert getattr(by_name["think"], "_strix_bounded", False) is True