Local‑first AI coding agent framework for terminal and editor workflows
Developers often need AI‑powered code assistance but worry about sending proprietary code to the cloud. This framework provides a runtime that powers both a CLI and VS Code/Cursor extension, letting the agent run locally with any Ollama model. It uses a skill‑based system to read, edit, test, and refactor code directly on the developer’s machine. By being model‑agnostic and privacy‑first, it offers more control than hosted AI coding tools.
View on GitHub →Burtson-Labs/bandit-agent-framework