| name | box-status |
| description | Report the box-memory workspace status - tier, capabilities, teams, memory counts, index health, recent activity. Read-only. Use when the user asks what's in their Box workspace, wants to show workspace status, asks how many memories or what tier, or wants a sanity check before a heavy operation. |
| argument-hint | [--refresh-tier] [--check-indexes] [--team=<name>] |
/box-status
If you see unfamiliar placeholders or need to check which tools are connected, see CONNECTORS.md.
Produce a human-readable snapshot of a box-memory workspace. Read-only — never modifies anything. Useful as a sanity check, an audit summary, or a pre-flight before a heavy operation.
Usage
/box-status [--refresh-tier] [--check-indexes] [--team=<name>]
Examples:
/box-status — standard report (cached tier info)/box-status --refresh-tier — re-probe Box capabilities before reporting/box-status --check-indexes — flag index drift suspicions/box-status --team=engineering — scope to one team
What to do
- Read
_box-memory.json. If missing, output "No box-memory workspace found at this Box location. Run /box-init to create one." Don't error.
- (If
--refresh-tier) Call box-tier-detect to re-probe and update cached capabilities.
- Read workspace-root rollup
_index.json for total counts. If missing or older than 24h, walk per-folder indexes. Note staleness in the report.
- Per-team stats. For each team in
_box-memory.json.teams[], read the team's memories/_index.json and count by status/kind. If a team folder is 403-inaccessible, mark "access denied" and continue.
- (If
--check-indexes) Compare each per-folder index's updated_at to the folder's modified_at from Box. Flag suspect folders. Don't rebuild — suggest /box-index-rebuild [folder].
- Compose report with sections: workspace info, account (tier + capabilities), routing, contents (totals by status/kind), teams, index health, most recent memories (top 5), top tags (top 5), settings, next-step commands.
- Flag actionable issues. Drift suspicions, tier-mismatch hints, stale-OAuth-token symptoms, missing compliance declarations.
Don't
- Don't modify anything. Status is read-only.
- Don't trigger a tier re-probe unless
--refresh-tier was asked — cached value is fine.
- Don't rebuild indexes during status; surface drift, suggest
box-index-rebuild.
- Don't load every memory's full content; index entries are enough.
Errors to surface clearly
- Box MCP not connected → standard MCP message.
- Permission denied reading workspace root → check Box folder permissions on the workspace root.
Deep reference
For the full report template, drift-detection logic, and actionable-issue heuristics:
