Hornet: A Local Coding Goal You Can Inspect, Pause, and Resume

Description and use

Hornet enables developers to persist coding goals, inspect results, and resume work after interruptions using a local model and a worker VM.

Problem

Developers need a durable coding workflow that preserves state, allows inspection, pausing, and resuming, and records execution evidence on disk.

Outcome

Hornet enables developers to define coding goals as cells, persist all execution evidence, and resume interrupted work, improving reliability and traceability of model‑driven code generation.

Workflow

• Prepare host: install Python 3.13, VirtualBox, LM Studio, OpenSSH, GnuPG.
• Download Hornet v0.15.1 starter ZIP, verify signature and SHA‑256 manifest.
• Extract starter and run PowerShell commands to create a fresh Python 3.13 virtual environment and install the wheel.
• Execute first‑run helper to create a new project, run diagnostics, and write a hello.txt contract file.
• Create a dedicated Ubuntu 22.04 cloud‑VMDK worker VM, verify its checksum, and provision it with Python, Git, and OpenSSH.
• Configure Hornet to use LM Studio on 127.0.0.1:1234 and the worker VM on 127.0.0.1:2224.
• Run Hornet CLI to repair a Python function, inspect the result, and observe checkpointing and resumption capabilities.

Evidence

Hornet was created and is led by Daniel F. MacDonald, with engineering contributions credited in the project.

Source

The first‑run helper creates a new project, runs local diagnostics, and writes a small hello.txt file.

Source

Hornet separates the saved work from the model’s current working context.

Source