39K views · 379 reactions | Comment “7” and I’ll send you all 7 GitHub links. If you build AI agents, stop rebuilding everything from scratch. These 7 repos give your agents some seriously useful capabilities: Browser Use — lets your agent operate a real browser, click through pages, and complete workflows. Agent Memory — gives your coding agent persistent memory so it can carry context across tasks. Scientific Agent Skills — 160+ skills designed to turn an AI agent into a powerful research assistant. Diagram Design — a collection of diagrams built specifically for Claude Code workflows. Anthropic Cybersecurity Skills — cybersecurity capabilities designed for modern coding agents. Awesome Harness Engineering — orchestration patterns and resources for building better agent systems. Open Viking — a context system for giving agents better access to the information they need. These can save you months of building agent infrastructure yourself. Comment “7” and I’ll send you all 7 links. Make sure you're following so the DM lands. #GitHub #AIAgents #AI #ClaudeCode #AItools | Duncan Rogoff
[0:00]These seven GitHub repos will save you months of building AI agents,
[0:03]and almost nobody is talking about them yet.
[0:05]So first up is browser use, where your agent opens a real browser
[0:08]and runs the whole workflow the way a person would,
[0:11]clicking through pages and filling everything out on its own.
[0:13]Then agent memory gives your coding agent memory that sticks,
[0:16]so it learns across tasks instead of starting from zero every second.
[0:20]Third is scientific agent skills, which is 160 plus skills