Adds safety mechanism to ensure only explicitly allowed tools can be executed.
When a model attempts to call a tool not in the allowed set (via `tools`
parameter or `tool_choice`), the call fails with an error message rather
than executing arbitrary code.

- Implements `allowed_tools()` to extract valid tool names from specs
- Updates `mk_toolres()` to accept `limit_to` parameter for filtering
- Modifies `Chat.__call__()` to validate tool calls automatically

- We use the same approach as claudette that uses limit_ns and relies on toolslm.func_call to throw KeyError
- Improved mk_tool_choice let user select "none", and if None is provided returns NOT_GIVEN istead of None.
This fixes with openrouter that gives 400 when tool_choice is null.
- Test were failing as we use `models[1]` to get the model name, which become `gpt-5.2-pro`.