What this does
The agent inventories its own current setup — tools, data sources, assumptions — and produces a ranked gap list: what it cannot see, what it is assuming without evidence, and what single failure would hurt most.
What to expect back
At least one HAVE-BUT-STALE you didn't expect, one assumption you'd never articulated, and a top gap whose fix costs an hour, not a subscription. Rerun monthly; keep the lists; diff them.
Works with
- gryps_capabilitiespublic server
- gryps_venue_snapshotpublic server
Agents fail silently at their edges. The operator who asks "what can't you see?" monthly stays ahead of the operator who asks "what do you think?" daily.
The prompt
next: s4-kill-your-favorite-idea
S · Sharpening