Skip to main content
An agent drives Smithers, not you. Ask your coding agent for an outcome; it runs the work, watches it, and clears approvals for you. Set up two ways: paste a prompt and let your agent wire everything, or run three commands yourself.

Fastest path: let your agent do it

Paste this into your harness (Claude Code, Codex, Cursor, Copilot, or any shell-capable agent). It installs the workflow pack, syncs Smithers’ generated CLI skill files where supported, registers the MCP server, and reports which starter fits your next task.
Setup prompt
init prints what it scaffolds, the agents it detected, and next commands to try. skills add syncs every Smithers-owned skill to supported locations, both the generated CLI command skills and the curated smithers skill, with the same behavior in agent and CI sessions as in a terminal; skills list reports the status of both. mcp add registers an MCP server for detected agents: Claude Code, Codex, Cursor, Gemini, OpenCode, Amp, Windsurf, and more.

Or wire it by hand

Scope the MCP install with --agent <name>, or the skills sync with --no-global to target the repo instead of home:
No skills directory? Point it at bunx smthrs docs-full (prints the full docs bundle) or bunx smthrs ask "<question>", and add a few lines to AGENTS.md or CLAUDE.md so it remembers Smithers is available.

Pick your harness

Each page covers exact paths, config files, and harness-specific steps. The full support matrix and three integration surfaces (skill, MCP server, standing instructions) live in Agent Support.

Verify it worked

Ask your agent to list the workflows; under the hood it runs:
If the MCP server registered, the agent can call list_workflows, run_workflow, watch_run, and resolve_approval. See the MCP Server reference for every tool.

Next steps