hummingbot-deploy — Claude Code skill: deploy Hummingbot
This skill walks an agent through standing up Hummingbot trading infrastructure: the API server, an MCP server and the Condor Telegram bot.
npx skills add hummingbot/skills --skill hummingbot-deploy -a claude-codeWhat hummingbot-deploy does
Hummingbot API is the part you cannot skip. It serves a REST API for trading, market data and running bot strategies across CEXs and DEXs. The MCP server and Condor are add-ons: the first lets an AI agent talk to that API, the second gives you a terminal and Telegram front end.
Before anything installs, the skill runs check_env.sh to look for Docker, Docker Compose, Git, Make, an interactive TTY and signs of a container. On a normal machine make setup followed by make deploy is enough. Inside a container there is no TTY, so the interactive prompts fail; the skill writes .env by hand and patches docker-compose.yml because bind mounts do not survive Docker-in-Docker. No Docker and Docker Compose means no Hummingbot API at all, so skipping the environment check wastes your time.
The MCP server goes in through install_mcp.sh with --agent, --user and --pass, reusing the credentials from the API setup. Once everything is up, restart your CLI so the MCP server loads. To upgrade later, run git pull and make deploy inside hummingbot-api.
What it helps with
- stand up Hummingbot API on a machine or in a container
- wire the MCP server into Claude, Gemini CLI or Codex CLI
- confirm Docker, Docker Compose, Git and Make are present
- run the Condor Telegram bot against the API
- upgrade an existing Hummingbot API install
Example prompts for the agent
- Deploy Hummingbot API and the MCP server with admin/admin credentials
- Check the environment and install Hummingbot in a container without a TTY
- Upgrade hummingbot-api and restart the services
How to install hummingbot-deploy
1Skills CLI (Vercel)
npx skills add hummingbot/skills --skill hummingbot-deploy -a claude-code2Manually: put the files in place
~/.claude/skills/hummingbot-deploy/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/hummingbot-deploy/
1Skills CLI (Vercel)
npx skills add hummingbot/skills --skill hummingbot-deploy -a codex2Manually: put the files in place
.agents/skills/hummingbot-deploy/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add hummingbot/skills --skill hummingbot-deploy -a cursor2Manually: put the files in place
.cursor/skills/hummingbot-deploy/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install hummingbot/skills hummingbot-deploy2Manually: put the files in place
.github/skills/hummingbot-deploy/SKILL.md1Skills CLI (Vercel)
npx skills add hummingbot/skills --skill hummingbot-deploy -a gemini-cli2Platform CLI
gemini skills install https://github.com/hummingbot/skills --path skills/hummingbot-deployManually: put the files in place
.devin/skills/hummingbot-deploy/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add hummingbot/skills --skill hummingbot-deploy -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about hummingbot-deploy
How do I install hummingbot-deploy in Claude Code?
Run npx skills add hummingbot/skills --skill hummingbot-deploy -a claude-code. The skill lands at ~/.claude/skills/hummingbot-deploy/SKILL.md.
What does hummingbot-deploy do?
It deploys Hummingbot API and, optionally, the MCP server and Condor. The skill checks the environment, writes .env and starts the services with make deploy.
What does the skill need to work?
Docker, Docker Compose, Git and Make, all verified by check_env.sh. In a container without a TTY the skill skips the interactive setup and creates .env itself.
