11 November 2026

Open source design system from Meta for React

Astryx is an open source design system that grew inside Meta over eight years, with 150+ accessible components, theming, dark mode, templates, and a CLI. People and AI assistants use the same tooling.

• Uses pre-built CSS and typed React components, with no build plugin or new styling library to adopt.
• A theme is a set of CSS custom property overrides, and swizzle ejects a component's source into your project.
• Requires React 19 or later, and is currently in Beta.

Official Agent Skills examples from Anthropic for Claude

This repo holds Anthropic's implementation of skills for Claude: folders of instructions, scripts, and resources that Claude loads for specialized tasks. Provided for demonstration and educational purposes.

• Each skill is self-contained in one folder with a SKILL.md file holding instructions and metadata.
• Examples range from art, music, and design to web app testing, MCP server generation, and enterprise workflows.
• The docx, pdf, pptx, and xlsx skills are included as a reference, under Apache 2.0.

Production-grade engineering skills for AI coding agents

A set of skills encoding the workflows, quality gates, and best practices senior engineers use, packaged so agents follow them across every phase of development.

• Nine slash commands follow the development lifecycle, such as /spec, /plan, /build, /test, /review, and /ship.
• /build auto generates the plan and implements every task after one approval, and each task is still tested and committed individually.
• Skills activate automatically based on the work, for example api-and-interface-design when designing an API.

Starter projects for building apps with the Claude API

Each quickstart gives a foundation you can build on for applications using the Claude API. You need an API key from console.anthropic.com.

• Examples include a Customer Support Agent with a knowledge base and a Financial Data Analyst with interactive data visualization via chat.
• There is a Computer Use Demo, a native-macOS Computer Use Best Practices project, and a Browser Use Demo.
• Best Practices covers prompt caching, server-side compaction, batched tool calls, a sandboxed shell, and trajectory recording.

Multi-joint dynamics with contact simulator from Google DeepMind

MuJoCo is a general purpose physics engine for research and development in robotics, biomechanics, graphics, animation, and machine learning. It is maintained by Google DeepMind.

• Has a C API, and the runtime simulation module works on low-level data preallocated by the built-in XML compiler.
• Interactive visualization is available in native desktop windows and web browsers, plus Python bindings and a Unity plug-in.
• Can be tried in the browser at mujoco.org with nothing to install.

Self-hosted personal finance app, a community fork of Maybe

Sure is a fully working personal finance app and a community fork of the abandoned Maybe Finance project. It can be self-hosted with Docker.

• Can be accessed from a browser, the macOS desktop app, the mobile app, API clients, and LLM agents.
• Licensed AGPLv3. The README asks forks to include the original license and state they are not affiliated with Maybe Finance Inc.

Configure and automate Home Assistant with AI assistants

ha-mcp lets Claude, ChatGPT, and other AI assistants configure and automate Home Assistant. The recommended way to run it is the HA-MCP integration.

• The integration installs through HACS or the Marketplace (Home Assistant 2026.11+) and runs the full server in-process on every installation type.
• Setup asks for an administrator access token, and the README suggests a dedicated account for HA-MCP.
• Breaking change in v7.3.0: ha_config_set_yaml moved to beta.

Browser automation CLI for AI agents, written in Rust

agent-browser is a native Rust CLI that gives an agent a browser it can drive one command at a time. snapshot returns the page as compact text with a ref such as @e1 for each element.

• The agent acts on a ref, reads the page again to check the result, and repeats.
• A background daemon controls the browser over the Chrome DevTools Protocol, and each --session gets its own daemon and browser.
• The browser stays open between commands, so each step is one short shell call.

Find social media accounts by username across 400+ networks

Sherlock hunts down accounts for a given username across more than 400 social networks. Found accounts are stored in a text file per username.

• Can search for one or several usernames at once, for example sherlock user1 user2 user3.
• Output options include CSV, xlsx, JSON, and text files, plus --proxy, --timeout, and --site.
• Installs with pipx install sherlock-project, or via pip, uv, and Docker.

Full-stack web framework for Rust, rendered on the server

Topcoat is a full-stack framework for Rust that prioritizes simplicity and productivity. It is early-stage and experimental, so expect breaking changes.

• All markup renders on the server, and async components can query the database directly with no separate API layer.
• A $(...) expression is type-checked Rust evaluated on the server and translated to JavaScript, with no wasm bundle or client build step.
• #[shard] re-renders a component on the server when its $(...) arguments change, then swaps the new HTML in place.

Context-compression plugin for month-long agent sessions

billion-context sits between an agent and its model API, rewriting Anthropic/OpenAI streams. It targets small context windows (100K is enough) and 5x fewer tokens.

• The README cites a healthy prefix-cache hit rate of 95-97%, with compression costing at most 2%.
• Its 1,500-item third-party evaluation ranks it #1 on ContextBench, and 79.20% vs 62.60% for the best RAG on LongMemEval-S.
• The paper and benchmark code are in the repo, with the paper under the MIT License.

Daily-use agent skills for engineers, from a .agents directory

A set of skills the author uses every day for real engineering, not vibe coding. They are small, easy to adapt, and composable, and work with any model.

• Install as a self-updating plugin or through skills.sh, which copies editable files into your project. Pick one per agent.
• Install steps are given for Claude Code, Codex, GitHub Copilot, Gemini CLI, and other agents such as cursor and opencode.
• When installing from skills.sh, the README says to include setup-matt-pocock-skills.