CodeNav: Beyond tool-use to using real-world codebases with LLM agents

TL;DR

CodeNav is an LLM agent that intelligently navigates and leverages unseen code repositories to solve complex user queries, surpassing traditional tool-use methods by automatically indexing, searching, and utilizing code snippets.

Contribution

It introduces a novel LLM agent that automatically interacts with real-world codebases, reducing manual tool registration and enhancing code understanding capabilities.

Findings

CodeNav effectively solves complex queries across diverse codebases.

It outperforms traditional tool-use LLM agents in benchmark tests.

Source code access improves agent performance more than natural language descriptions.

Abstract

We present CodeNav, an LLM agent that navigates and leverages previously unseen code repositories to solve user queries. In contrast to tool-use LLM agents that require ``registration'' of all relevant tools via manual descriptions within the LLM context, CodeNav automatically indexes and searches over code blocks in the target codebase, finds relevant code snippets, imports them, and uses them to iteratively generate a solution with execution feedback. To highlight the core-capabilities of CodeNav, we first showcase three case studies where we use CodeNav for solving complex user queries using three diverse codebases. Next, on three benchmarks, we quantitatively compare the effectiveness of code-use (which only has access to the target codebase) to tool-use (which has privileged access to all tool names and descriptions). Finally, we study the effect of varying kinds of tool and library descriptions on code-use performance, as well as investigate the advantage of the agent seeing source code as opposed to natural descriptions of code. All code will be made open source under a permissive license.