eth-node_skill

Overview

This skill manages Ethereum execution client nodes so you can start, stop, monitor sync, inspect peers, and view logs. It supports common EL clients like reth and geth and assumes a local RPC at http://localhost:8545 by default. The goal is fast operational guidance for healthy, secure node operation.

How this skill works

The skill issues shell commands and JSON-RPC calls to the node's HTTP endpoint to control and inspect the client. It can start nodes with safe localhost bindings, stop processes, query eth_syncing and admin_* RPC methods for sync and peer details, and recommend log tails and diagnostics. It emphasizes local-only admin and security best practices.

When to use it

• Start or stop a local reth or geth execution node with safe RPC binding.
• Check whether a node is fully synced or in progress (eth_syncing).
• List, add, or inspect connected peers using admin_peers/admin_addPeer.
• Gather node info, chain ID, or network version via JSON-RPC.
• Diagnose stuck syncs, zero peers, or missing RPC responses.

Best practices

• Bind RPC to 127.0.0.1 and never expose admin/debug namespaces on public interfaces.
• Use JWT auth for Engine API when running an execution client with a consensus client.
• Redirect background process logs to files and use tail -f for live inspection.
• Check peer counts and eth_syncing first for a quick health summary.
• Open discovery ports (30303 TCP/UDP) only behind a firewall and monitor disk I/O and free space.

Example use cases

• Start reth locally with: reth node --http --http.addr 127.0.0.1 --http.api eth,net,web3 &> reth.log 2>&1 &
• Confirm node fully synced with eth_syncing JSON-RPC; false means synced.
• List peers to troubleshoot network isolation: call admin_peers on localhost RPC.
• Resolve stuck sync by checking iostat, df -h, and restarting the client with debug options.
• Add a trusted peer manually using admin_addPeer with an enode URL when bootnodes are unreachable.