Lord Claude

Use OpenAI's GPT, xAI's Grok and Google's Gemini as Claude Code agents.

Lord adds three agent types: lord:gpt, lord:grok and lord:gemini. Claude works with them the way it works with its own subagents. It can start them in the background, message them, stop them, give them a worktree, or fan them out in a Workflow. Each agent runs in its vendor's own command-line tool, on your own subscription. Claude never answers their turns, so their work costs no Claude tokens. Their descriptions add about 600 tokens to each session.

Install

claude plugin install lord --marketplace aylee1024/lord-claude

Then start a new Claude Code session. You need:

• Claude Code 2.1.302 or later
• Python 3.11 or later (Lord looks in Homebrew, /usr/local/bin and your PATH; set LORD_PYTHON to pick one)
• macOS (Linux is untested)
• at least one vendor tool, installed and signed in:

Lord uses your subscription logins. It never reads login files, never uses API keys, and removes secret-looking environment variables before it starts a vendor tool.

Use

Ask in plain words: "have gpt fix this test in a worktree, and have grok review it read-only." Claude picks the agent and the settings.

You can also put options on their own lines at the start of a task or message:

Where each agent can write: gpt in its folder, @dir folders and temp folders; grok in its folder, temp folders and ~/.grok; gemini anywhere, because agy has no sandbox. A sandbox limits one agent, not the agents it starts, so use @readonly when an agent must not change anything. A read-only agent cannot start or message other agents.

Type /lord to see your agents, the model and effort each ran with, their token use, and each agent's activity log.

How it works

The plugin answers each Lord agent's turn by passing it to lordd.py, a small background program on your computer. That program runs one vendor tool per agent, keeps each agent's conversation for later messages, and stops the tool when you stop the agent. It keeps running after you close Claude Code, so a long task finishes and its result arrives with your next message to that agent. It exits after 10 idle minutes. After you update Lord, the old program exits as soon as its agents are idle, and the next agent starts the new one.

The default models are gpt-6.1-sol, grok-4.7 and gemini-4-argon. If your account does not offer one, or the vendor retires it, Lord uses the vendor tool's own default model instead; /lord shows the model each agent ran. A model you name with @model is used as named, or refused with the list your account offers. You can change default models and limits in ~/.claude/lord/config.json, for example {"grok": {"model": "grok-4.7"}}.

What Lord runs, sends and stores

• Sends. Your task goes only to the vendor tool you chose, which sends it to that vendor under your account. Lord makes no network requests of its own. The plugin code talks only to the background program, over a local socket at ~/.claude/lord/lordd.sock. See PRIVACY.md.
• Runs. lordd.py with Python, /usr/bin/curl to read answers from the local socket, /bin/ps every few seconds while an agent works (for the memory limit), and the vendor tools you installed. It never starts a shell and never downloads anything. The install commands above are for you to run; Lord does not run them.
• Reads. Only the turns of Lord agents; your main conversation and other agents pass through untouched. From the environment it reads HOME, PATH, LORD_PYTHON and LORD_DAEMON_SOCKET.
• Hooks. It answers Lord agents' turns (turn.start, turn.step), notes new Lord agents and their folders (agent.spawn, SubagentStart), hands a Lord agent's tool results back to its vendor (tool.call, inputs unchanged), delivers a message to a Lord agent that is mid-turn (session.send, session.receive), stops a vendor's turn when Claude stops the agent (turn.complete, session.end), and adds /lord and a status line (session.start, command.run). It changes no settings, permissions or tool inputs.
• Stores in ~/.claude/lord/: the socket; a random key file the plugin and background program share, which grok's sandbox cannot read; one line per agent with its settings and token counts (never task text); each agent's activity log, with secrets removed; and the program's own log. Anything unused for 14 days is removed.
• Changes two files elsewhere. While the background program runs, agy's settings hold an entry named lord, so Gemini agents get Lord's tools; your own agy sessions started in that time also start it, and it offers them nothing. The program removes the entry when it exits. It adds two sandbox profiles to ~/.grok/sandbox.toml, between # lord: begin and # lord: end, and leaves your own profiles alone. Lord uses its own profiles because grok runs unprotected when one of its built-in profiles fails, while a custom profile that fails stops grok from starting.

CONTRACT.md has the full technical details.

Uninstall

claude plugin uninstall lord@lord-claude
pkill -f lordd.py         # stops Lord's background program, if it is running
agy mcp remove lord       # only if you used Gemini agents
rm -rf ~/.claude/lord     # Lord's records and logs

Lord's grok profiles stay in ~/.grok/sandbox.toml; delete the lines from # lord: begin to # lord: end to remove them.

Older version

Version 1 was a set of skills (/codex, /gemini, /grok, /composer, /review-panel). It is kept at the v1-skills tag.

Development

claude plugin validate .
claude plugin test .
python3 -m pytest tests -q

License

MIT