patt_describe_data
ChatGPTDiscover event, ticket, report, anonymous-record, named-client and single-contact datasets; learn their record units, relationships, intended tool routes, consent requirements, current capabilities and known omissions. Read before choosing a dataset or metric. Anonymous output or missing consent never proves names are absent from PATT. This describes the bounded MCP contract, not raw database fields or server configuration.
patt_get_access_context
ChatGPTRead the connected account and available read-only capabilities. Use first; permissions are resolved from current PATT records.
patt_get_client_history
ChatGPTRequires patt:clients:read consent and current PATT CRM permissions. Use an opaque clientRef returned by patt_list_clients for this connected account. Read authorized current ticket entries and sold table bookings for the selected events, with attendance filtering and separately defined per-currency operational values. No contacts or raw profile IDs. Event dates are event starts, not proof of visits. Follow nextCursor with identical selection and filters.
patt_get_client_profile
ChatGPTRequires both patt:clients:read and patt:contacts:read consent plus current PATT contact permissions. Read one client name and permitted email/phone using an opaque clientRef from this account and an authorized event selection. A successful null field is unavailable in the permitted source; a denied request is an error, not evidence that the stored contact is empty. Optional includeBirthday additionally requires patt:birthdays:read; returns only month/day with source status, never year, and never treats a defaulted date as verified. This does not export contacts, expose raw profile IDs, or modify PATT.
patt_get_combined_performance_report
ChatGPTGenerate the seven-sheet combined Report-tab Excel in memory for current MCP permissions. Select up to 20 eventIds, or a dateFrom/dateTo range of at most 366 days containing at most 50 permitted events; clubIds optionally narrows it. All supplied filters intersect. Use patt_get_report_options to discover clubs/events and ask missing selection questions. With no event/date selection returns needs_selection, never an unbounded export. Customer names, emails and demographics are always omitted, even with CRM consent; report shares never widen MCP permissions. This is a privacy-preserving report, not an identical customer-detail export. Returns an embedded XLSX resource plus summary/coverage, with no upload, storage write or hosted URL. Save the resource as the given filename if the client supports files. Oversized reports fail completely; narrow selection instead of claiming partial totals.
patt_get_event_details
ChatGPTRead event description, venue, artists, age/dress rules, currencies and booking slots, plus stored presentation, checkout requirements/limits and table booking/selection configuration. Current account event visibility is enforced. Slot times use local wall-clock strings; absent legacy switches are null. Configuration is not effective availability, eligibility, table inventory, attendee details or sales.
patt_get_event_summary
ChatGPTRead ticket holdings and scanned ticket-entry counts for an accessible event. Owner/organizer or team scope follows PATT permissions. Counts include free/bonus entries and exclude tables and checkout holds. This does not report paid sales, money, or unique people.
patt_get_report_options
ChatGPTDiscover only permitted report clubs and events before preparing a combined tickets-and-tables Excel. Ask which clubs, the date range (including timezone), and all matching events or selected events. Omit clubIds for all permitted clubs; supplied clubs and date bounds intersect. Follow nextCursor for events; choices grant no access and export rechecks everything. Options contain no customer data. A client may display the returned choices in its own UI or ask conversationally.
patt_list_clients
ChatGPTRead permitted client names with literal case-insensitive name query, kind and attendance filters. Requires patt:clients:read consent AND current CRM permissions. Event IDs OR a bounded UTC event-start range; attendance filters matching activity BEFORE identity grouping and counts. An identity may have both arrived and not-arrived activity; unknown attendance matches neither filter. Profiles and separate unlinked sold bookings are not verified unique people or every dashboard seat alias. Filter grouped identities by count ranges, first/last event-start dates and an explicitly named money metric/currency; sort by those values. querySummary totals cover every match before pagination. clientOperationalValue is configured ticket face value plus table Stats revenue, not captured or net spending. Birthday month/day filtering requires patt:birthdays:read and reports unverified/default-possible sources. Check selection, querySummary, unnamed coverage and every nextCursor before claiming all matches. No contacts, profile IDs or admission credentials; use separate authorized profile lookup when needed.
patt_list_event_records
ChatGPTRead anonymous operational ticket-entry or table-booking units, filtered by attendance, type, category or booking state. For participant/client names or name search use patt_list_clients instead: this dataset deliberately omits names even when CRM names are permitted. Null displayName is a legacy placeholder, never proof names are missing or forbidden. Reporting and CRM permissions differ. No contacts, holder IDs, payment IDs or admission credentials. Follow nextCursor with identical filters.
patt_list_event_ticket_types
ChatGPTRead discoverable ticket types with configured prices, sale windows, per-order limits, required checkout fields, pay-at-door/offline settings, waitlist switches, pack tiers and phase/display settings. Private ticket discovery follows the account permissions separately from reporting. These are stored settings, not live stock, current phase state, a final quote or purchase eligibility. Legacy missing settings are null. Follow nextCursor for the same event.
patt_list_events
ChatGPTFind accessible events by literal name, club, storage status and time. For last/latest already-started events set period=past, sort=newest; for next events set period=upcoming, sort=oldest. period=all (default) includes future dates, even distant ones. past means start <= current server time, not necessarily finished. Archived is storage, not past/future. Optional UTC date bounds intersect the period. Response selection states the clock used; every page refreshes current time and permissions. Follow nextCursor with the same filters and sort. Use IDs for event details, metrics, entries or named clients.
patt_query_event_reports
ChatGPTRead broad revenue_overview or attendance_overview presets, OR exact metrics, for event IDs or a bounded UTC event-start range. Revenue overview separates tickets/tables, online/offline/door and combined totals; never add overlapping rows. Ranking needs one explicit sort.metric and money currency. If intent is unclear ask the user, or return the structured needs_clarification question; never silently choose table-only revenue. The server may request a supported clarification form. Current permissions apply independently per section; inspect component scopes, exclusions, quality and coverage. Booked value and configured Event Stats revenue are not settled payments or profit. Shared-event mirrored tickets are excluded from direct-sales totals.
patt_query_team_performance
ChatGPTCompare and rank authorized promoter, canonical manager credit-owner or commercial-team performance across selected events. Select explicit metrics; group by entity and optionally break down by event or club, filter a permitted entity reference/name, and page results. Money ranking requires one currency. Optional ticketFilter selects ticket type, attendance and evidence-based payment channel only for ticket metrics. Current per-event ticket/table and identity-visibility rules apply independently. Private team totals do not grant access to their roster; management does not transfer sales ownership; commercial teams may overlap and must not be summed. Definitions use Event Stats operational values, not the legacy ranking, captured-payment or commission-settlement basis. Inspect definitions, coverage, quality and groupingDefinition. Ask for missing period, measure or currency when it changes the answer. No customer/contact data or raw staff IDs are returned.