Agent Intelligence Platform

describe_intelligence (metadata) and list_intelligence (inventory) are clearly distinct, but query_intelligence and search_intelligence heavily overlap — both involve locating answers on shelves, one by probing a known shelf and one by keyword discovery. An agent could easily pick the wrong one when it just wants to find a relevant answer.

All four tools follow a strict verb_intelligence pattern (describe_, list_, query_, search_), which is predictable and easy to scan. No mixed conventions or camelCase deviations.

Four tools is reasonable for a discovery-and-metadata layer, though it sits near the thin end given the platform also involves publication and payment flows. Each tool does earn a distinct role, so it is not gratuitously padded.

The surface covers discovery (list, search, describe, query) but exposes no publish/create-shelf operation despite advertising 'publication status,' and purchase is only reachable via an external x402 HTTP client. This leaves a notable gap between the read-only discovery layer and the implied authoring/selling lifecycle.