Showed that fine-tuning a GPT model on public GitHub code yields a capable program synthesizer, and introduced HumanEval — the docstring-to-function benchmark that still anchors code-generation evaluation. A production variant powers GitHub Copilot.
AI Client2023
Adds agent-native UI patterns to apps through chat, generative UI, shared state, human-in-the-loop flows, and AG-UI-based frontend integrations.
MCP Server2025
Lets AI agents drive GitHub in natural language via MCP: browse repos, triage issues, review pull requests, and trigger Actions runs. Runs as a GitHub-hosted remote OAuth server or a local Go binary, with per-toolset scoping and a read-only mode.
MCP Client2025
Bridges LLM-driven AI assistants to the Unity Editor so models can create scenes, edit C# scripts, manage assets, run tests and automate game-dev workflows. Exposes 47 focused MCP tool entrypoints, supports many MCP clients, and is MIT-licensed for local use.
MCP Server2025
Exposes Azure DevOps resources (projects, repos, pipelines, work items, wikis) to AI agents via an MCP server; remote-first (hosted HTTP endpoint) with an optional local stdio Node.js server for VS Code and domain-filtered toolsets.
AI Client2025
Source code for the GitHub Copilot Chat extension in VS Code: inline chat, an agent mode that plans and edits files autonomously, next-edit suggestions, and MCP integration. Open-sourced so developers can study how Copilot connects to an editor.
MCP Server2025
Helps developers connect IDEs and AI agents to the Figma MCP server to extract design context, generate code from frames, and write updates back to Figma. Includes client setup (VS Code, Cursor, Claude Code, Gemini CLI), tools and skills, best practices, and beta rate-limit notes.
AI Client2025
Transforms enterprise architecture governance into a structured, AI-assisted workflow covering principles, requirements, risk, procurement and design reviews. Bundles templates, CLI/plugins and autonomous research agents (MCP integrations) to keep traceability and compliance.
AI Coding2025
Desktop app that runs many CLI coding agents — Claude Code, Codex, Cursor, Gemini — in parallel, each in its own git worktree and branch. A built-in diff viewer, terminal, and PR tracking let you dispatch and review 10+ agents at once.
AI Agent2025
Curated index of 1,100+ official and community Agent Skills for extending AI agents across platforms — Claude Code, Codex, Gemini CLI, Cursor, GitHub Copilot, OpenCode, and more. Focuses on hand-picked, engineering-authored skills, organized by vendor, paths, and quality guidance.
AI Coding2025
Provides adaptive workflow steering rules for AI coding agents to guide development across Inception, Construction, and Operations phases. Includes opt-in extensions (security, testing), IDE/agent integrations (Cursor, Kiro, Amazon Q, Copilot, Claude), and human-in-the-loop approval points.
AI Client2025
Drives AI coding agents through a five-phase loop — discuss, plan, execute, verify, ship — offloading heavy work to fresh-context subagents to fight context rot. The main session stays lean while parallel waves do the building.