20+ skills and MCP servers bridging AI and blockchain development

AI is quickly moving from something developers experiment with to something they build on top of, and nowhere is that shift more visible right now than in blockchain development.

Just a few years ago, large language models entered the mainstream. ChatGPT launched in late 2022, followed by Claude, Grok, and Gemini through 2023, while developer tools like Cursor began embedding AI directly into coding environments. Since then, the pace of breakthroughs with this emerging technology has accelerated rapidly.

“Vibecoding” is already beginning to change how developers build products, and that shift is likely to continue as AI becomes more embedded in developer workflows.

Many in the web3 community have started publishing official MCP servers and skills for LLMs like Claude, including chains like Solana, that help expand developer understanding and interaction with their ecosystems.

“Official” here means the skill or MCP server is published or maintained by a protocol team, core developers, or a recognized ecosystem partner, while “unofficial” refers to community-built tools created independently by developers without formal endorsement from the underlying project.

Here we highlight twenty plus MCP servers and skills for developers to explore. Turnkey doesn’t endorse any of these, or vouch for their security, but we align with the broader goal of experimentation and adoption of AI skills across the crypto ecosystem.

How do skills and MCP servers differ?

While both MCP servers and skills expand what AI systems can do, they operate at different layers.

MCP servers (Model Context Protocol servers) are essentially infrastructure connectors and structured interfaces that allow AI models to access external tools, data, APIs, and operational systems in a consistent way. They expose blockchain nodes, analytics platforms, smart contract tooling, security services, or other infrastructure so AI can interact directly with them. 

In web3, this often means structured access to smart contract workflows, blockchain data, or production infrastructure. Put simply, MCP servers help AI connect to blockchain infrastructure and make those systems more accessible to AI-driven workflows.

Skills, on the other hand, focus on workflow customization. They package prompts, instructions, scripts, or resources that help AI assistants perform specific tasks more consistently once they have context. 

Rather than providing new infrastructure access, skills shape how AI behaves and executes tasks. Typically markdown-based, they allow developers and “vibe coders” to tailor AI platforms like Claude and others for focused workflows.

 In February 2026, Anthropic released The Complete Guide to Building Skills with Claude, a resource outlining how developers can build and implement skills within their own projects. Put simply, skills help AI make practical use of the access MCP servers provide.

A quick security reality check

As exciting as this space is, there are already signs that AI skill ecosystems may face some of the same security challenges long seen in open-source package registries. 

Researchers at 1Password recently documented how agent skill marketplaces can become an attack surface, specifically highlighting issues in the OpenClaw ecosystem. In one case, a widely downloaded AI skill for Twitter that appeared to be a normal productivity tool was found distributing infostealing malware through what looked like routine setup instructions.

‍

Further research found that out of 2,857 skills identified on OpenClaw’s marketplace ClawHub, there were 341 malicious entries, many designed to harvest login credentials, crypto wallet keys, SSH access, and other sensitive data.

The takeaway is not to avoid experimentation. It is simply to treat AI skills and integrations with the same caution you would apply to any developer dependency:

• Review the source code carefully
  
  
• Understand what permissions or data access a skill requires
  
  
• Avoid running untrusted skills on sensitive machines or environments
  
  
• Assume “top downloaded” does not automatically mean safe
  
  

This is especially relevant in crypto environments, where credentials, signing keys, API access, and automation workflows can carry real financial risk if compromised.

20+ MPC servers and skills built for onchain workflows

Below is a snapshot of AI skills and MCP servers helping shape how onchain workflows get built and operate. Some focus on security and auditing, others expose blockchain infrastructure, analytics, or developer tooling.

Together, they show how AI is moving from simply analyzing blockchain data to actively interacting with onchain systems.

Security, auditing, and risk analysis

These focus on smart contract security, threat detection, and audit workflows.

Trail of Bits Claude Skills
https://github.com/trailofbits/skills
AI skills for security research, vulnerability detection, and smart contract audit workflows.

Slither MCP (Trail of Bits)
https://github.com/trailofbits/slither-mcp
MCP server exposing Slither static analysis tools for Solidity smart contract security.

OpenZeppelin MCP
https://github.com/OpenZeppelin/openzeppelin-mcp
MCP servers connecting AI agents to OpenZeppelin smart contract security and development tooling.

GoPlus Security MCP
https://github.com/GoPlusSecurity/goplus-mcp
Web3 security data MCP server for risk analysis, malicious contract detection, and transaction safety.

Core blockchain infrastructure and node access

These expose chain infrastructure directly to AI agents.

Alchemy MCP Server
https://github.com/alchemyplatform/alchemy-mcp-server
Official MCP server for interacting with Alchemy blockchain APIs and infrastructure services.

Chainstack RPC Nodes MCP
https://github.com/chainstacklabs/rpc-nodes-mcp
MCP interface for interacting with blockchain JSON-RPC nodes across multiple networks.

Nodit MCP Server
https://github.com/noditlabs/nodit-mcp-server
Multi-chain blockchain data infrastructure MCP server for indexing and analytics access.

Helius MCP Server (dcSpark)
https://github.com/dcSpark/mcp-server-helius
Infrastructure MCP server providing structured Solana blockchain data access.

Protocol-specific MCP servers

Focused on enabling AI interaction with specific ecosystems.

Base MCP Server (Coinbase ecosystem)
https://github.com/base/base-mcp
Official Base L2 ecosystem MCP server enabling AI-driven onchain interactions.

NEAR MCP
https://github.com/nearai/near-mcp
AI tooling MCP server providing access to NEAR blockchain infrastructure.

Solana MCP (SendAI / Agent Kit)
https://github.com/sendaifun/solana-mcp
Agent-focused MCP server for building AI systems interacting with Solana.

Market data, analytics, and DeFi intelligence

These give AI access to onchain analytics and trading intelligence.

DEXPaprika MCP (Coinpaprika)
https://github.com/coinpaprika/dexpaprika-mcp
DEX trading analytics MCP server providing liquidity, pricing, and market data.

CoinStats MCP Server
https://github.com/CoinStatsHQ/coinstats-mcp
Official MCP server for crypto portfolio tracking, analytics, and market data.

Dune Analytics MCP
https://github.com/kukapay/dune-analytics-mcp
Community MCP server exposing Dune analytics data to AI agents.

DeFi Llama MCP (dcSpark)
https://github.com/dcSpark/mcp-server-defillama
Structured DeFi ecosystem analytics and TVL data MCP server.

Developer skills and AI workflow tooling

More workflow-oriented AI capabilities rather than raw infrastructure.

Solana Developer Skill (Solana Foundation)
https://github.com/solana-foundation/solana-dev-skill
Claude skill focused on Solana development workflows and ecosystem navigation.

QuickNode Solana Anchor Claude Skill
https://github.com/quiknode-labs/solana-anchor-claude-skill
Developer skill for building Solana programs using Anchor with QuickNode.

Monad Claude Web3 Plugin Marketplace
https://github.com/monad-developers/claude-web3-plugin-marketplace
Collection of experimental Web3 AI plugins curated around the Monad ecosystem.

Trail of Bits Security Skills
https://github.com/trailofbits/skills
Security-focused developer workflow skills for AI-assisted audits.

Community or unofficial AI skills and MCP projects

Useful signals of ecosystem experimentation even if not officially supported.

StartWithBitcoin Skill
https://github.com/bramkanstein/startwithbitcoin-skill
Educational AI skill focused on Bitcoin fundamentals and onboarding.

Arkham Intelligence Claude Skill
https://github.com/Vyntral/arkham-intelligence-claude-skill
AI skill for blockchain intelligence workflows using Arkham-style analytics.

Hashgraph Online Tools
https://github.com/hashgraph-online
Collection of Hedera-related AI tools and integrations.

Arbitrum dApp Skill
https://hummusonrails.github.io/arbitrum-dapp-skill/
Community-built skill for interacting with Arbitrum decentralized applications.

Turnkey: At the center of onchain agentic security

AI and blockchain infrastructure are beginning to intersect in practical ways. MCP servers expose protocol capabilities directly to AI systems, while skills package workflows that help developers actually use those capabilities. 

Platforms like Turnkey are focusing specifically on the security layer of this shift, providing policy-based wallet infrastructure, secure signing environments, verifiable infrastructure and builds and programmable controls designed to support AI-driven onchain workflows without sacrificing operational safety.

It’s still early. Standards are evolving. Security practices are catching up. But the direction is clear. AI is not just analyzing blockchain data anymore. It’s starting to interact with blockchain infrastructure directly.

For builders in web3, this creates both opportunity and responsibility. The opportunity is faster development, automation, and new forms of programmable finance. The responsibility is making sure those integrations are secure, verifiable, and designed for real-world use.

Experiment freely, but experiment securely.

Start building with Turnkey today. 

‍