[memory refactor][3/n] Introduce RAGToolRuntime as a specialized sub-protocol (#832)

See https://github.com/meta-llama/llama-stack/issues/827 for the broader
design.

Third part:
- we need to make `tool_runtime.rag_tool.query_context()` and
`tool_runtime.rag_tool.insert_documents()` methods work smoothly with
complete type safety. To that end, we introduce a sub-resource path
`tool-runtime/rag-tool/` and make changes to the resolver to make things
work.
- the PR updates the agents implementation to directly call these typed
APIs for memory accesses rather than going through the complex, untyped
"invoke_tool" API. the code looks much nicer and simpler (expectedly.)
- there are a number of hacks in the server resolver implementation
still, we will live with some and fix some

Note that we must make sure the client SDKs are able to handle this
subresource complexity also. Stainless has support for subresources, so
this should be possible but beware.

## Test Plan

Our RAG test is sad (doesn't actually test for actual RAG output) but I
verified that the implementation works. I will work on fixing the RAG
test afterwards.

```bash
pytest -s -v tests/agents/test_agents.py -k "rag and together" --safety-shield=meta-llama/Llama-Guard-3-8B
```

llama_stack/distribution/stack.py

@@ -29,7 +29,7 @@ from llama_stack.apis.scoring_functions import ScoringFunctions
 from llama_stack.apis.shields import Shields
 from llama_stack.apis.synthetic_data_generation import SyntheticDataGeneration
 from llama_stack.apis.telemetry import Telemetry
-from llama_stack.apis.tools import ToolGroups, ToolRuntime
+from llama_stack.apis.tools import RAGToolRuntime, ToolGroups, ToolRuntime
 from llama_stack.apis.vector_dbs import VectorDBs
 from llama_stack.apis.vector_io import VectorIO
 from llama_stack.distribution.datatypes import StackRunConfig
@@ -62,6 +62,7 @@ class LlamaStack(
     Inspect,
     ToolGroups,
     ToolRuntime,
+    RAGToolRuntime,
 ):
     pass