Eth2 Quickstart

Eth2 Quickstart

Automates production-ready Ethereum node setup on Linux servers — execution client, consensus client, MEV integration, OS hardening, and optional RPC exposure. Supports 7 execution clients (Geth, Besu, Nethermind, Erigon, Reth, Nimbus-eth1, Ethrex) and 6 consensus clients (Prysm, Lighthouse, Teku, Nimbus, Lodestar, Grandine).

This skill is repo-aware: it runs inside an eth2-quickstart checkout and routes all actions through ./scripts/eth2qs.sh. It never invents new lifecycle commands.

Install

# Recommended
clawhub install eth2-quickstart

# Or clone directly
git clone --depth 1 https://github.com/chimera-defi/eth2-quickstart.git
cd eth2-quickstart

What agents can do with this skill

• Bootstrap a fresh host: detect the next safe install step and execute it
• Check node health: machine-readable JSON output from doctor --json
• Operate services: start, stop, restart, view logs
• Update clients: update-all covers all installed components
• Safe cleanup: clean-data --dry-run shows what would be removed before any deletion; secrets and validator keystores are always preserved

Commands

# Detect the next safe install step (machine-readable)
./scripts/eth2qs.sh plan --json

# Preview next step without executing
./scripts/eth2qs.sh ensure

# Execute next step (requires explicit confirmation)
./scripts/eth2qs.sh ensure --apply --confirm

# Install Ethereum node (execution + consensus + MEV)
sudo ./scripts/eth2qs.sh phase1
./scripts/eth2qs.sh phase2 --execution=geth --consensus=prysm --mev=mev-boost

# Health check
./scripts/eth2qs.sh doctor --json

# Operate
./scripts/eth2qs.sh start
./scripts/eth2qs.sh stop
./scripts/eth2qs.sh logs --run2 -n 200
./scripts/eth2qs.sh stats

# Cleanup (always dry-run first)
./scripts/eth2qs.sh clean-data --dry-run
./scripts/eth2qs.sh clean-data --confirm
sudo ./scripts/eth2qs.sh cleanup-host --dry-run

Safety guardrails

• ensure --apply requires explicit --confirm — no silent destructive execution
• clean-data preserves ~/secrets, validator keystores, and wallet directories by design
• doctor --json detects service-unit drift (running binary doesn't match unit file)
• Agents must require human confirmation before phase1 (root/reboot), destructive cleanup, or host-wide changes
• Agents must never generate validator keys or remove secrets

doctor --json output shape

{
  "summary": { "passed": 12, "warnings": 2, "failed": 0, "status": "warn" },
  "checks": [
    { "status": "pass", "name": "RAM: 32GB (recommended: 16GB+)", "details": "" },
    { "status": "warn", "name": "Execution client (eth1): Not installed", "details": "" }
  ]
}

Use doctor --json as the canonical agent-readable health surface. Use status / stats / logs for human-readable RCA.

Minimum host requirements