I build AI-native tools, agent systems, and developer workflows.
My current focus is on making software agents more structured, observable, and useful in real engineering work: harnesses, multi-agent coordination, browser/runtime automation, writing tools, and knowledge workflows.
I care about tools that make thinking, writing, coding, and agent collaboration more controllable.
| Project | What it is | Stack |
|---|---|---|
| Open Agent Harness | A protocol-driven runtime for coding agents, focused on structured tools, explicit agent contracts, and observable multi-agent collaboration. | TypeScript / Bun |
| Markly | A WYSIWYG Markdown editor extension for VS Code with rich editing, table support, image handling, PDF/HTML export, and diagnostics. | JavaScript / VS Code |
| Lumen | A native macOS PDF reader designed for researchers, students, and power users. | Swift / SwiftUI |
| Light Finger Browser | Browser fingerprint, proxy, browser instance, and stealth runtime orchestration infrastructure. | Go |
| Agent Team | A multi-agent project management system with role-based agents, task scheduling, tools, memory, observability, and plugin architecture. | TypeScript |
- Agent harness design
- Multi-agent coordination
- AI coding workflows
- Browser automation and runtime control
- Markdown, PDF, and knowledge tools
- Developer tools for thinking, writing, and building
- agentlab.fun - AI learning and research notes
- GitHub
- Email: goodjin2@gmail.com
