Cron, systemd, MCP servers, skills. Vitals reads what is already running on your machine and hands your agent a health check before it touches any of it.
No dashboard, no telemetry leaving the box. Every command reads local session logs and system state, then reports back to the agent that asked.
The rest do the plumbing: install, detect, mcp, tokens, coach, sessions, compact, snapshot.
That is not a CLI you read. That is a tool your agent calls from inside its own loop, before it does anything risky.
Detects pi, Claude Code, Cursor, OpenCode, and Codex CLI, and wires the MCP server plus a priming skill into whichever ones you run.
Restart your agent host afterward. The next session calls vitals_summary on its own before anything risky.