Yes. Open-source under the MIT license. Your AI tool may have its own costs, but the skill itself is free.

Always verify. The skill gives your AI accurate reference material, but it can still hallucinate contract addresses, invent function signatures, or produce incorrect code. Treat AI output the same way you'd treat code from a junior developer: review it, test it, and check addresses against official sources before deploying anything on-chain.

Complete contract references (ABIs with descriptions), subgraph schemas and query patterns, SDK guides, architecture documentation, and helper scripts for resolving addresses, token lists, and balances.

Without the skill, your AI relies on training data that may be outdated or wrong. With the skill, it reads verified, current contract interfaces and integration guides at the moment it needs them. The difference shows up most in integration code, debugging, and anything involving specific function signatures or contract addresses.

Just run the standard npx skills add superfluid-org/skills install. It writes to the .codex folder, which the Codex Desktop App picks up automatically. No extra steps needed.

Run npx skills update to pull the latest version. For ClawHub, run npx clawhub@latest install superfluid again.

The skill works with any agent that supports the Skills protocol: Claude Code, OpenCode, Cursor, GitHub Copilot, Cline, Windsurf, Gemini CLI, Codex CLI, Codex Desktop App, and more. Claude Desktop is supported via the plugin marketplace or ZIP upload.

The skill has been most tested with Claude Code using the Opus model. Results will vary across different models and tools. A more capable model with extended thinking (or equivalent reasoning mode) will make better use of the references. Other tools and models are compatible, but that combination is the most proven.