MCP
get_schema
Discover item types, fields, writable target types, and the read-query catalog before doing anything else.
Call get_schema first. Every other tool call keys off the field keys, target types, and query signatures it returns.
| Field | Type | Required | Notes |
|---|---|---|---|
type |
string | No | A target type or item type slug to inspect. Omit to get the full discovery payload. |
get_schema requires the read:data scope.
What You Get Back
Section titled “What You Get Back”What comes back depends on the type argument you send.
No arguments. Returns the full discovery payload:
queryReference: the catalog of every read query field and its signature, the same catalog documented in full on GraphQL.itemTypes: summaries of the tenant’s configured item types.targetTypes: writable non-item targets for suggest_change, each with its allowed operations and fields.linkingTypes: the linking targets:itemrelationship,stepitemlink,stepdocumentlink.
{ "name": "get_schema", "arguments": {}}type set to a target type. Pass a value such as "step", "workflow", or "dashboard" to get that target’s schema: name, description, allowed operations, and field definitions.
{ "name": "get_schema", "arguments": { "type": "step" }}type set to an item type slug. Pass a value such as "issue" to get the full item type, including its field definitions and select option values.
{ "name": "get_schema", "arguments": { "type": "issue" }}Unknown type. Returns a text response listing the valid target and linking types and suggesting a bare get_schema call.
Agent Notes
Section titled “Agent Notes”- Use the field keys and option values
get_schemareturns exactly: don’t infer them from labels. - Slugs are tenant-specific. Never reuse a slug from one tenant on another.
- Configuration changes over time. Re-check
get_schemaafter an administrator edits item types or fields.