Dynamic response_format with create_agent

### Checked other resources - [x] This is a feature request, not a bug report or usage question. - [x] I added a clear and descriptive title that summarizes the feature request. - [x] I used the GitHub search to find a similar feature request and didn't find it. - [x] I checked the LangChain documentation and API reference to see if this feature already exists. - [x] This is not related to the langchain-community package. ### Package (Required) - [x] langchain - [ ] langchain-openai - [ ] langchain-anthropic - [ ] langchain-classic - [ ] langchain-core - [ ] langchain-cli - [ ] langchain-model-profiles - [ ] langchain-tests - [ ] langchain-text-splitters - [ ] langchain-chroma - [ ] langchain-deepseek - [ ] langchain-exa - [ ] langchain-fireworks - [ ] langchain-groq - [ ] langchain-huggingface - [ ] langchain-mistralai - [ ] langchain-nomic - [ ] langchain-ollama - [ ] langchain-perplexity - [ ] langchain-prompty - [ ] langchain-qdrant - [ ] langchain-xai - [ ] Other / not sure / general ### Feature Description Support callable/functional response_format parameters in create_agent() that resolve the schema at runtime based on agent state, rather than requiring static schemas bound at agent creation time. Currently, response_format must be a static schema (Pydantic model, TypedDict, or JSON schema dict) that is known when the agent is created. This limitation prevents use cases where the response structure needs to vary dynamically based on: User preferences stored in state, Conversation context, etc. ### Use Case Agent that dynamically constrains a Pydantic model's `Literal` or `Enum` field based on conversation state: ```python def get_response_format(state: dict, runtime) -> type[BaseModel]: """Return schema with dynamically constrained options.""" available_actions = state.get("available_actions", ["read", "write"]) # Dynamically create a model with Literal constrained to available actions class DynamicResponse(BaseModel): action: Li…