# Dynamic ReAct: Scalable Tool Selection for Large-Scale MCP Environments

arXiv: https://arxiv.org/abs/2509.20386 · HTML: https://mlsanity.com/paper/2509.20386/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 7 references that are in the map.
- First posted on arXiv 22 Sep 2025.

## Builds on (most cited first, 3 of 3)

- [ScaleMCP: Dynamic and Auto-Synchronizing Model Context Protocol Tools for LLM Agents](https://mlsanity.com/paper/2505.06416/) (2025)
- [MCP-Zero: Active Tool Discovery for Autonomous LLM Agents](https://mlsanity.com/paper/2506.01056/) (2025)
- [RAG-MCP: Mitigating Prompt Bloat in LLM Tool Selection via Retrieval-Augmented Generation](https://mlsanity.com/paper/2505.03275/) (2025)

## Built on by (most cited first, 6 of 6)

- [CodeMem: Architecting Reproducible Agents via Dynamic MCP and Procedural Memory](https://mlsanity.com/paper/2512.15813/) (2025)
- [What Does a Skill Actually Do? Estimands and Evaluation Validity for Tool and Skill Use in LLM Agents: A Critical Review](https://mlsanity.com/paper/2609.33153/) (2026)
- [Synthesizing Procedural Memory: Challenges and Architectures in Automated Workflow Generation](https://mlsanity.com/paper/2512.20278/) (2025)
- [Trie Automata for Constrained Decoding over Large Finite Sets](https://mlsanity.com/paper/2608.12574/) (2026)
- [Toollery: Scaling LLM Agents to Thousands of Skills and Tools](https://mlsanity.com/paper/2609.22218/) (2026)
- [What Can Component-Replacement Evidence Establish? A Critical Scoping Review of Local Decisions in LLM Agents](https://mlsanity.com/paper/2609.39989/) (2026)

## Abstract (arXiv)

We present Dynamic ReAct, a novel approach for enabling ReAct agents to efficiently operate with extensive Model Control Protocol (MCP) tool sets that exceed the contextual memory limitations of large language models. Our approach addresses the fundamental challenge of tool selection in environments containing hundreds or thousands of available tools, where loading all tools simultaneously is computationally infeasible. We propose and evaluate five distinct architectures that progressively refine the tool selection process, culminating in a search-and-load mechanism that achieves intelligent tool selection with minimal computational overhead. Our experimental results demonstrate that the proposed approach reduces tool loading by up to 50% while maintaining task completion accuracy, advancing the path towards truly general-purpose AI agents capable of dynamically adapting to diverse task environments.