Skip to content

Latest commit

 

History

History
 
 

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

README.md

Blockscout Analysis

AI agent skill for blockchain data analysis and building tools that query on-chain data through the Blockscout MCP Server. Works across all EVM chains indexed by Blockscout.

What this skill does

The skill gives an agent domain expertise in two areas:

  1. Blockchain activity analysis — answering questions about wallets, tokens, transactions, contracts, and NFTs using the Blockscout MCP Server.
  2. Building scripts and tools — writing and executing code that queries blockchain data via the MCP REST API for deterministic, multi-step flows.

Key ideas

  • Single data source: All data access goes through the Blockscout MCP Server (native MCP tools or REST API). Chainscout is used only to resolve chain IDs to Blockscout explorer URLs.
  • Data source priority: Dedicated MCP tools first (enriched, LLM-friendly), then direct_api_call for endpoints not covered by dedicated tools.
  • Execution strategy: The skill guides the agent to choose between direct tool calls, scripts (REST API), hybrid approaches, and LLM reasoning — based on task complexity and whether deterministic logic or semantic understanding is needed.
  • Multichain: Almost all MCP tools accept a chain_id parameter. The agent uses get_chains_list to discover supported chains.

Directory structure

blockscout-analysis/
├── SKILL.md                           # Agent entry point — decision framework,
│                                      #   workflow, and all behavioral instructions
├── README.md                          # This file (human overview)
└── references/                        # Lookup data consulted during execution
    ├── blockscout-api-index.md        # Endpoint index for direct_api_call
    ├── blockscout-api/                # Detailed endpoint parameter references
    │   ├── addresses.md
    │   ├── blocks.md
    │   ├── tokens.md
    │   ├── transactions.md
    │   ├── smart-contracts.md
    │   ├── ...                        # + chain-specific files (arbitrum, zksync, etc.)
    │   └── stats.md
    └── chainscout-api.md              # Chain registry endpoint
  • SKILL.md is the self-contained agent entry point. An agent that reads only this file has everything it needs to behave correctly.
  • references/ contains API endpoint details the agent looks up during execution (e.g., to find parameters for direct_api_call).

Setup

Integration depends on your agent platform. For optimal results, configure the Blockscout MCP server (https://mcp.blockscout.com/mcp) in your agent's MCP settings. The skill can fall back to the REST API if native MCP is not available.

License

MIT