Problem

A first-time user following the Quick Start reads this in the README:

"Call the mem_status tool" → confirms the server is connected

So they learn mem_status as the health-check primitive. But mem_status is the MCP tool name, not a CLI command. There's no equivalent mm status in the CLI:

$ mm status
Usage: mm [OPTIONS] COMMAND [ARGS]...
Try 'mm -h' for help.

Error: No such command 'status'.

Looking at mm --help, none of the existing commands map cleanly to "is my install healthy / what's in the DB":

• mm config show — configuration, not runtime state
• mm watchdog status — periodic-check results, not current inventory
• mm recall / mm search — requires a query
• (no simple mm status / mm info / mm inventory)

Meanwhile mem_status (MCP) returns exactly what a user wants for a 5-second sanity check:

memtomem Status
==============
Storage:   sqlite
DB path:   /Users/pdstudio/.memtomem/memtomem.db
Embedding: none /
Dimension: 0
Top-K:     10
RRF k:     60

Index stats
-----------
Total chunks:  3
Source files:  3
...

Suggested direction

Add mm status as a thin CLI wrapper over the same function mem_status MCP tool calls. Benefits:

• Symmetric naming between MCP (mem_status) and CLI (mm status) — easier to teach
• Gives a post-install verification step that (a) proves the binary works, (b) proves config is readable, (c) shows the populated DB — without requiring a query
• Fills the gap between mm init ("I set it up") and mm search ("I actually use it")

A more conservative alternative: document the correct incantation (mm config show && sqlite3 ~/.memtomem/memtomem.db "SELECT COUNT(*) FROM chunks" or similar) in Troubleshooting, so users don't guess mm status.

Repro context

Fresh-install smoke test, 2026-04-22, memtomem==0.1.23.