rilo (CLI)

One command to connect AI agents to Rilo:

npx rilogtm connect

Interactive multi-select over detected clients, then wires the Rilo MCP endpoint into each one:

| Client | How it's wired | | --- | --- | | Claude Code | claude mcp add --scope user --transport http rilo <url> | | Cursor | ~/.cursor/mcp.json → mcpServers.rilo.url | | Codex | codex mcp add rilo --url <url>, falls back to ~/.codex/config.toml | | OpenClaw | ~/.mcporter/mcporter.json → mcpServers.rilo.baseUrl | | Claude.ai / Desktop | manual — URL copied to clipboard + 3-step guide |

The CLI never handles auth or tokens. The client triggers Rilo's OAuth on the first tool call; the backend auto-provisions the user + org at that moment, so OAuth consent doubles as signup. LinkedIn connect happens in-chat via the connect_linkedin_account operation.

Flags

npx rilogtm connect --client claude-code,cursor --yes   # non-interactive
npx rilogtm connect --client all --yes                  # every detected client
npx rilogtm connect --dry-run                           # print plan, change nothing
npx rilogtm connect --url https://other-host/mcp        # override endpoint

Local development

cd cli && npm install
node cli.js connect --dry-run

Publishing checklist

1. Confirm the production MCP URL in cli.js (DEFAULT_URL) — move to https://mcp.getrilo.com/mcp once the custom domain exists.
2. npm login as the org owner.
3. npm publish (first publish claims the rilo name — also consider publishing rilogtm / rilomcp as alias packages that just depend on rilo, so the names are held).
4. Smoke-test from a clean machine: npx rilogtm@latest connect --dry-run.