ChainLaunch

Claude Code Skills

ChainLaunch ships with Claude Code skills — structured markdown files that teach AI coding assistants how to operate ChainLaunch from natural language.…

ChainLaunch ships with Claude Code skills — structured markdown files that teach AI coding assistants how to operate ChainLaunch from natural language. Skills work in Claude Code, OpenAI Codex CLI, Cursor, and Gemini CLI.

With skills installed, you can type things like:

  • "Create a Fabric testnet with 2 orgs and 3 orderers"
  • "Deploy a Besu network with 4 validators"
  • "Review my Go changes for ChainLaunch patterns"

And the AI assistant will execute the correct ChainLaunch CLI commands, API calls, and verification steps automatically.

Available Skills

ChainLaunch maintains 6 skills across the monorepo:

Skill Location Purpose
chaindeploy-cli chaindeploy/.claude/skills/ Full CLI reference for ChainLaunch Core — all commands, flags, and workflows
chainlaunch-pro-cli chainlaunch-pro/.claude/skills/ CLI reference for ChainLaunch Pro — includes enterprise commands (install, update, encrypt, sharing)
deploy-chainlaunch .claude/skills/ Server deployment guide — VPS provisioning, Docker setup, TLS configuration
go-review .claude/skills/ Go code review checklist — error handling, sqlc patterns, RBAC, API conventions
write-changelog .claude/skills/ Changelog generation from git commits using Conventional Commits format
technical-docs .claude/skills/ Documentation generation orchestrator for the full docs system

Installing Skills

Quick Install with skills.sh

The fastest way to add ChainLaunch skills to your project is using skills.sh — the open agent skills ecosystem. Skills installed this way work across Claude Code, GitHub Copilot, Cursor, Cline, and other AI coding tools.

Install all ChainLaunch skills at once:

npx skills add LF-Decentralized-Trust-labs/chainlaunch --all

Install a specific skill only:

npx skills add LF-Decentralized-Trust-labs/chainlaunch --skill chaindeploy-cli

Install globally (available in all projects):

npx skills add LF-Decentralized-Trust-labs/chainlaunch --all --global

List available skills before installing:

npx skills add LF-Decentralized-Trust-labs/chainlaunch --list

Other skills commands

# List installed skills
npx skills list
 
# Check for updates
npx skills check
 
# Update all skills to latest
npx skills update
 
# Remove a skill
npx skills remove chaindeploy-cli

Alternative: Clone the repository

If you prefer to work directly from the source, clone the ChainLaunch repository. Skills are automatically loaded when you open the repo in Claude Code:

git clone https://github.com/LF-Decentralized-Trust-labs/chainlaunch.git
cd chainlaunch
claude

Skills live in .claude/skills/ directories and are picked up by the AI assistant automatically.

Prerequisites

  • Claude Code (or another compatible AI coding tool)
  • Node.js (for npx skills commands)

Using Skills

Example: Create a Fabric Network

Open a terminal in the ChainLaunch repo and start Claude Code:

claude

Then type:

Create a Fabric testnet called "supply-chain" with org SupplyOrg,
2 peers, and 3 orderers

Claude Code reads the chaindeploy-cli skill and runs:

chainlaunch testnet fabric --name supply-chain --org SupplyOrg \
  --peerOrgs SupplyOrg --ordererOrgs Orderer1 \
  --peerCounts SupplyOrg=2 --ordererCounts Orderer1=3 \
  --provider-id 1

It then automatically verifies nodes are running, sets anchor peers, and pulls the connection profile.

Example: Deploy a Besu Network

Spin up a 4-node Besu testnet with QBFT consensus

Claude Code runs:

chainlaunch testnet besu --name my-besu --nodes 4 --mode service

Example: Deploy ChainLaunch to a Server

Deploy ChainLaunch to my server at 203.0.113.10

The deploy-chainlaunch skill guides the AI through SSH setup, Docker installation, binary download, TLS configuration, and service registration.

Skill Reference

chaindeploy-cli

The core skill covering the full ChainLaunch CLI. Key commands:

# Start the server
chainlaunch serve --port 8100 --data ~/.chainlaunch --dev
 
# Create a Fabric testnet (orgs, nodes, network, channel join — all in one)
chainlaunch testnet fabric --name mynet --org Org1 \
  --peerOrgs Org1 --ordererOrgs Orderer1 \
  --peerCounts Org1=2 --ordererCounts Orderer1=3
 
# Create a Besu testnet (QBFT consensus, min 4 nodes)
chainlaunch testnet besu --name mybesu --nodes 4
 
# Pull a Fabric connection profile
chainlaunch fabric network-config pull \
  --network mynet --msp-id Org1MSP \
  --output connection-profile.yaml \
  --url http://localhost:8100/api/v1 --username admin --password admin123
 
# Install chaincode (full lifecycle in one command)
chainlaunch fabric install \
  --chaincode mycc --channel mychannel \
  --policy "OR('Org1MSP.member')" \
  --chaincodeAddress "127.0.0.1:9999" \
  --config connection-profile.yaml \
  -u admin -o Org1MSP --local
 
# Invoke and query chaincode
chainlaunch fabric invoke --mspID Org1MSP --user admin \
  --config connection-profile.yaml \
  --channel mychannel --chaincode mycc --fcn Transfer -a arg1 -a arg2
 
chainlaunch fabric query --mspID Org1MSP --user admin \
  --config connection-profile.yaml \
  --channel mychannel --chaincode mycc --fcn GetAsset -a arg1

chainlaunch-pro-cli

Extends the core CLI with enterprise commands:

# Install as a system service (systemd/launchd)
chainlaunch install --name chainlaunch-pro \
  --executable /usr/local/bin/chainlaunch-pro --port 8100
 
# Self-update from GitHub releases
chainlaunch update --tag v1.2.0
 
# Encrypt/decrypt data
chainlaunch encrypt "sensitive-data" --key "$KEY_ENCRYPTION_KEY"
chainlaunch decrypt "encrypted-string" --key "$KEY_ENCRYPTION_KEY"
 
# Inspect encrypted database values
chainlaunch crypto decrypt-db --db chainlaunch.db \
  --table vault_configs --column token --id 1

go-review

Invoked with /go-review. Reviews Go code changes against ChainLaunch patterns:

  • Error handling conventions
  • SQLite/sqlc query patterns
  • RBAC permission enforcement
  • API handler structure
  • Test coverage

write-changelog

Invoked with /write-changelog. Parses git commits (Conventional Commits format) and generates a structured changelog entry grouped by features, fixes, and breaking changes.

deploy-chainlaunch

Invoked with /deploy-chainlaunch. Guides deployment to a production server:

  1. SSH into target server
  2. Install prerequisites (Docker, unzip, Java for Fabric)
  3. Download and configure ChainLaunch binary
  4. Set up TLS certificates
  5. Register as a system service
  6. Verify health check

Fabric Network Lifecycle

When using skills to create a Fabric network, the AI follows this complete lifecycle:

  1. Create network — chainlaunch testnet fabric provisions orgs, nodes, network, and joins all peers/orderers
  2. Verify nodes — Checks all nodes show RUNNING status via the API
  3. Set anchor peers — Configures anchor peers for cross-org gossip (REST API call)
  4. Pull connection profile — Generates the network config YAML for client applications
  5. Install chaincode — Full lifecycle: package, install, approve, commit
  6. Test — Invoke and query chaincode to verify the network is operational

The skill ensures no step is skipped. Without anchor peers, for example, chaincodes can't discover peers from other organizations.

Creating Custom Skills

You can add your own skills for project-specific workflows. Create a SKILL.md file in .claude/skills/<skill-name>/:

.claude/
  skills/
    my-custom-skill/
      SKILL.md

The SKILL.md file uses this format:

---
name: my-custom-skill
description: >
  Brief description of what this skill does and when to use it.
---
 
# Skill Title
 
Instructions, commands, patterns, and examples that teach
the AI assistant how to perform this workflow.

Skills are version-controlled with your codebase, so they stay in sync with your tooling and evolve as your project does.

Further Reading