[Feature Request] Native Support for MCP Code Execution (Programmatic Tool Calling) & Lazy Result Loading

### 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 - [x] langchain-openai - [x] langchain-anthropic - [ ] langchain-classic - [x] 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 I am requesting native support for **Programmatic Tool Calling (Code Execution**) with Model Context Protocol (MCP) integration. Currently, LangChain agents typically operate in a "chatty" loop: 1. LLM predicts a tool call (JSON). 2. Runtime executes the tool. 3. Full result is appended to the context. 4. LLM reads the result and decides the next step. **The requested feature changes this paradigm to:** 1. The Agent is aware of available tools as importable libraries. 2. The Agent generates a script (e.g., Python or TypeScript) that orchestrates multiple tool calls, loops, and data filtering logic in one go. 3. The script runs in a sandbox (e.g., Docker, Daytona, E2B). 4. **Crucially**: Intermediate results stay in the sandbox variable state and are not automatically dumped into the LLM context unless explicitly requested or returned. **Key capabilities needed:** - **Batching via Code**: Ability for the LLM to write a script calling `tool_a()` and `tool_b()`…