hashgraph-online/awesome-ai-plugins

What is this project?

Awesome AI Plugins is a community-maintained catalog of extensions for AI assistants. It collects plugins, skills, MCP servers, apps, and agent tools that work with popular AI coding assistants like Codex, ChatGPT, Claude Code, Gemini CLI, Grok, Kimi, DeepSeek Harness, Cursor, OpenCode, and others.

Think of it as a curated directory — a place to discover what extensions exist, what they do, and where to get them. It is not an installer; each project has its own setup instructions.

What does it contain?

The catalog is organized into sections:

• Start Here – A quick workflow for building and submitting your own plugin, including a recommended validation step.
• Official Plugins – Plugins curated by OpenAI (like Box, GitHub, Slack, etc.) and a few others.
• Community Plugins – Extensions built by the community, split into categories like Agent Skills (individual installable skill collections), Grok Plugins, Kimi Plugins, and DeepSeek Harness Plugins.
• Formats & Development – Resources about formats and building your own extensions.
• Guides & Articles – Educational material.
• Related Projects – Other relevant tools.
• Claim Your Plugin – How to add your own project to the list.
• Plugin Trust Scores and Plugin Quality – Information about how projects are evaluated.
• Contributing – Guidelines for adding entries.

How does it work?

The README itself is the human-readable catalog. There are also machine-readable exports (plugins.json and .agents/plugins/marketplace.json) for automation and registry consumers. You can browse the list directly on GitHub or use the searchable HOL Plugin Registry at hol.org/registry/plugins.

Security and trust

The project encourages plugin authors to run a local plugin-scanner tool (via pipx run plugin-scanner lint . and verify .) and to add a scanner GitHub Action to their repository. This scanner checks for security issues and keeps the catalog safer for everyone.

Projects that maintain scanner CI receive a full trust score. Projects without it are still eligible but get a 10% reduction in their trust score. The catalog itself is independently scanned by HOL (the maintainers).

How to contribute

If you have a plugin, skill, or tool, you can submit it via a pull request (PRs are welcome). The README points to a CONTRIBUTING.md file for specific requirements.

Key takeaway

This is a discovery catalog — a well-organized, community-vetted list of AI assistant extensions. It helps you find tools, understand what they do, and get a sense of their trustworthiness, but you always install and use each extension according to its own project's instructions.