12 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.

LLM-powered stock analysis that runs on a schedule

A stock analysis system that combines an LLM, market data from several sources, and news search into a daily decision dashboard. It can run on a schedule through GitHub Actions without a server.

• Covers A-share, Hong Kong, US, Japan, Korea, and Taiwan stock markets.
• Results are pushed automatically to WeCom, Feishu, Telegram, Discord, Slack, or email.
• Free data sources such as AkShare, Baostock, and YFinance are built in, but their stability is not guaranteed.

Modern web platform guidance that steers coding agents

Source repo for Modern Web Guidance, a set of skills that puts web platform best practices and browser compatibility data into coding agents. It aims to steer agents away from legacy patterns.

• The CLI has search, retrieve, and legacy-scan commands to find guides and scan a codebase for old patterns.
• Stewarded by the Google Chrome, Microsoft Edge, Mozilla Firefox, and MDN teams.

Agent skill that writes plans for cheaper models to run

An agent skill that audits a codebase with your most capable model and writes implementation plans as markdown files. It never implements anything itself; a cheaper agent does the execution.

• Commands such as /improve quick, /improve security, and /improve branch set the audit scope.
• Plans land in plans/, one file each, and any agent can pick them up.
• /improve execute 001 runs a cheaper model in an isolated worktree and reviews its diff; merging stays up to you.

Engineering skills for AI coding agents

A collection of skills that encode the workflows and quality gates senior engineers use, packaged for AI agents across every phase from spec to ship. Nine slash commands map to the development lifecycle.

• Commands: /spec, /plan, /build, /test, /review, /ship, plus /constraints, /webperf, and /code-simplify.
• /build auto generates the plan and runs every task after one approval, with each task tested and committed on its own.
• One command installs all 25 skills into 70+ agents through the skills CLI.

Rust CLI that lets AI agents drive a browser

agent-browser is a native Rust CLI that gives an AI agent a browser to 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 the browser stays open between commands.
• Chrome for Testing is the default, and it can also attach to a running Chrome with --cdp.

Personal finance app you can self-host

Sure is a personal finance app, a community fork of the abandoned Maybe Finance project. It can be self-hosted with Docker and used from a browser, the macOS app, the mobile app, API clients, and LLM agents.

• Built from the Maybe Finance code, which was open sourced after the team stopped development in mid-2023.
• Licensed AGPLv3; other forks must include the original license and may not use the "Maybe" name or logo.
• A public dashboard shows slow requests from the demo site along with their stacktraces.

Configure and automate Home Assistant through AI assistants

ha-mcp is an MCP server for configuring and automating Home Assistant with Claude, ChatGPT, and other AI assistants. The recommended setup is the HA-MCP integration, which runs in-process inside Home Assistant.

• Installs through HACS, or through the Marketplace on Home Assistant 2026.11 and later; requires Home Assistant 2026.8.0 or newer.
• The server acts as an administrator account; the README suggests creating a dedicated HA-MCP account instead of using your own login.
• Breaking change in v7.3.0: ha_config_set_yaml moved to beta.

OpenAI CLI and SDK for finding and fixing security vulnerabilities

Codex Security helps find, investigate, and fix vulnerabilities in your code. Use it in Codex, through the @openai/codex-security CLI, or automate it with the TypeScript SDK and GitHub Action.

• Scans repositories, selected paths, or Git changes; validates candidate findings, generates patches, and verifies fixes.
• Findings can be tracked in Linear, Jira, or GitHub, and exported as SARIF, JSON, and CSV.
• Requires Node.js 22.13.0+ (22.x), 24.x, or 26.x, and Python 3.10+.

Algorithms implemented in Java for learning

A repo of algorithms implemented in Java for educational purposes. The README notes the implementations may be less efficient than the Java standard library.

• The full list of algorithms is in DIRECTORY.md.
• The code can be run and edited through Gitpod.io with one click.
• Contribution guidelines are in CONTRIBUTING.md and should be read before contributing.