AI Tooling

AI agents are powerful when they have context and useless when they don’t. The projects here are about closing that context gap — not by automating the coding itself, but by making sure the agent understands the codebase, the developer, and the actual task before it proposes anything.

OpenAgent

OpenAgent is a repo-aware CLI that reads your codebase, interviews you once via SOUL.md, and generates structured directives for AI coding agents. The directives are personalized to your architecture and style, not generic templates.

It is intentionally advisory: you review the generated directive, you decide what gets implemented, you maintain control. It’s live on PyPI as openagent-directive.

The Pattern

The problem isn’t that AI can’t write code. The problem is that every session starts from zero context. These tools fix the context layer: actual file structure, actual imports, actual conventions, and a persistent profile of how the developer thinks about the project.

Built with Python. Live on PyPI. The agent reads your repo before it writes anything.