minghinmatthewlam · GitHub
Skip to content
View minghinmatthewlam's full-sized avatar
📚
learning...
📚
learning...
63 followers · 0 following

Achievements

Achievement: Starstruckx3Achievement: Pair Extraordinairex4Achievement: Pull Sharkx3Achievement: YOLOAchievement: QuickdrawAchievement: Arctic Code Vault Contributor

Achievements

Achievement: Starstruckx3Achievement: Pair Extraordinairex4Achievement: Pull Sharkx3Achievement: YOLOAchievement: QuickdrawAchievement: Arctic Code Vault Contributor

Highlights

  • Pro

Block or report minghinmatthewlam

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
minghinmatthewlam/README.md

hey, i'm matthew 👋

product-minded engineer building ai products — agent harnesses, computer use, and evals

what i'm focused on

  • agent harnesses & dev tools — building the run loops, tooling, and interfaces that make ai agents actually useful
  • computer use & evals — giving agents hands to operate real apps, and rigorous ways to measure what works
  • product sense, 0→1 — end-to-end ownership from concept and design through architecture to polished ux

things i've built

project description
pi-gui desktop harness for the pi coding agent — electron + react, real-time streaming timeline with inline diffs, worktrees, and published npm extensions (goal mode + subagents)
computer-use-mcp agent-agnostic computer use as an mcp server — a signed swift binary that lets any agent see and operate mac apps in the background, without stealing cursor or focus
openbench eval harness comparing coding-agent harnesses — same model, harness varies: how much does the scaffolding around a model actually matter? checker-graded, reproducible
openci package manager and discovery platform for ai-agent ci workflows — find, install, and maintain github actions from any repo
release bots · @codexreleases · @pichangelog ai changelog bots watching github releases and auto-drafting threads with the claude api — ~20k followers combined
pluto 0→1 trading platform on hyperliquid — perps, spot, cross-chain deposits

stack

typescript python go react swift

𝕏 · @codexreleases · @pichangelog · linkedin · github

Pinned Loading

  1. pi-gui pi-gui Public

    Electron GUI app for the pi coding agent runtime

    TypeScript 1.3k 191

  2. computer-use-mcp computer-use-mcp Public

    Open, agent-agnostic MCP server exposing macOS computer use to any MCP client (Claude Code, Cursor, Codex, …). Background-safe, works on every app.

    Swift 50 4

  3. openbench openbench Public

    Same model, different wrapper: a from-scratch benchmark comparing coding-agent harnesses (codex, pi, opencode, cursor, devin) and open models on correctness, speed, and token cost

    Python 138 10

  4. openci openci Public

    Open-source CLI for installing AI-powered GitHub Actions workflows

    TypeScript 5

  5. agent-guards agent-guards Public

    Shared guardrails, skills, and bootstrap scripts for AI coding agents

    JavaScript 46 3