LitterBox+: An Extensible Framework for LLM-enhanced Scratch Static Code Analysis

2025-09-15

AI Summary

Large language models (LLMs) struggle to directly analyze or assist with block-based programming environments (e.g., Scratch) due to their non-textual, structural representation. Method: This paper introduces the first scalable framework that deeply integrates LLMs with static code analysis, embedded in the open-source tool LitterBox. It precisely parses Scratch block structures and converts them into semantically faithful textual representations—enabling LLM comprehension—and provides a lightweight API supporting multi-LLM orchestration, dynamic query processing, and generative repair code synthesis, all seamlessly integrated into the Scratch interface. Contribution/Results: The framework establishes the first LLM-driven static analysis and generative feedback loop for block-based programs. Its modular architecture supports diverse LLM backends and customizable analysis tasks. Empirical evaluation demonstrates its effectiveness and usability in code quality diagnostics and pedagogical support.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large language models (LLMs) have become an essential tool to support developers using traditional text-based programming languages, but the graphical notation of the block-based Scratch programming environment inhibits the use of LLMs. To overcome this limitation, we propose the LitterBox+ framework that extends the Scratch static code analysis tool LitterBox with the generative abilities of LLMs. By converting block-based code to a textual representation suitable for LLMs, LitterBox+ allows users to query LLMs about their programs, about quality issues reported by LitterBox, and it allows generating code fixes. Besides offering a programmatic API for these functionalities, LitterBox+ also extends the Scratch user interface to make these functionalities available directly in the environment familiar to learners. The framework is designed to be easily extensible with other prompts, LLM providers, and new features combining the program analysis capabilities of LitterBox with the generative features of LLMs. We provide a screencast demonstrating the tool at https://youtu.be/RZ6E0xgrIgQ.

Problem

• Extending Scratch static analysis with LLMs for code queries
• Converting block-based Scratch code to textual LLM representation
• Integrating LLM-generated fixes into Scratch programming environment

Innovation

• Extends Scratch analysis with LLM integration
• Converts block code to text for LLM queries
• Provides API and UI for Scratch environment