This section separates primary projects I build or actively maintain from related fork-based or participation-oriented work. Each topic lives in its own directory with paired English and Chinese READMEs.
| Project | Role | Problem | Stack | Stage |
|---|---|---|---|---|
| Token Insight / Repo | Builder / Maintainer | Make AI coding token usage inspectable without pushing local data to SaaS first | Rust, React, SQLite | Active build |
| Codex Composer / Repo | Builder / Maintainer | Turn ad hoc Codex usage into a repository workflow with explicit contracts and checks | JavaScript, Shell, Markdown | Active iteration |
| ChatGPT TurboRender / Repo | Builder / Maintainer | Keep long ChatGPT conversations responsive without replacing the native UI | TypeScript, WXT, Manifest V3, Playwright, Vitest | Active build |
| AgentScan / Repo | Contributor / Maintainer | Discover, fingerprint, and assess exposed AI agent instances across real networks | Go, React, SQLite | Active security tooling |
- Token Insight: local-first token observability for AI coding tools.
- Codex Composer: repository bootstrap for repeatable Codex workflows.
- ChatGPT TurboRender: browser-side long-thread responsiveness for ChatGPT.
- AgentScan: AI agent exposure discovery and security audit platform.
These repositories are included for context, but I do not present them as first-party projects built entirely from scratch inside DevLab.
| Project | Relationship | Notes |
|---|---|---|
| one-api | Fork / maintenance-oriented work | Relevant to LLM gateway, key distribution, and multi-provider API operations |
| MetaGPT | Fork / participation-oriented work | Relevant to multi-agent software workflow exploration |