Lab
These are open-source projects, experiments, and tools from my workflow. The pages are intentionally practical: understand the boundary, clone the repository, and see how it works without waiting for a product tour.
One ticket. Validated changes. Readable commits. One reviewable PR.
SWE Forge is a portable, opt-in, Git-native workflow layer for the coding agent you already use. It inspects the repository, chooses the smallest useful execution setup, implements and validates cohesive slices, runs a fresh independent review, and delivers one inspectable PR with a readable commit history. It never merges for you.
Think with context you own.
SelfContext keeps the context you choose in portable Markdown and makes it available to the AI tool you already use, so useful conversations can continue from your history, goals, decisions, constraints, and evidence instead of starting from zero.
A practical reference for developer tools, AI agents, and server utilities.
Developer Tools is a living collection of setup notes, configuration, and practical references for the tools around my development workflow. It covers AI coding agents, local model runtimes, developer environment utilities, and server infrastructure without pretending there is one perfect stack.
Reusable capabilities for OpenCode, Claude Code, and compatible AI coding agents.
This repository collects focused skills that give AI coding agents reusable guidance for engineering, validation, and review work. Each skill is a small, portable package with a clear purpose, so it can be installed when useful instead of becoming part of every prompt.
A small local-first desktop app for understanding AI coding workflows.
Nano Harness is a personal, pre-release Electron app for working with AI providers under controlled autonomy. It connects provider adapters, workspace tools, approvals, runs, persistence, and inspectable evidence in one local desktop experiment.
A practical path for running open-source models on Apple Silicon.
Local AI is a practical toolkit and notebook for running open-source language models on Apple Silicon with different local inference servers. It keeps the common workflow separate from runtime-specific instructions and machine-specific settings, so a working setup can be measured rather than guessed.