List all collections in the Directus instance

Get the schema for a specific collection

Get items from a collection with optional filtering and pagination

Create a new collection with optional fields. Creates a real table with an auto-increment id when no fields are given; pass folder: true for a grouping element instead.

Delete a collection (requires confirmation)

Create a new item in a collection

Update an existing item in a collection. In versioned collections the published item is read-only — target a draft version instead.

Delete items from a collection, either by explicit IDs or by a query. Providing neither deletes nothing.

Create a new field in a collection

Update an existing field in a collection

Delete a field from a collection (requires confirmation)

Execute bulk create, update, and delete operations

Analyze collection schema with relationship mapping and validation

Analyze relationships across collections

Create relationships between collections (O2O, O2M, M2O, M2M, M2A)

Validate collection schema and relationships

Diagnose collection access issues and permissions

Refresh Directus collection cache and verify access

Validate that a newly created collection is properly accessible

Get all users with optional filtering and pagination

Get a specific user by ID

Get files with optional filtering and pagination

Get flows with optional filtering and pagination

Get a specific flow by ID with optional operations

Create a new automation flow

Update an existing flow

Delete a flow and all its operations (requires confirmation)

Manually trigger a flow execution

Get flow operations with optional filtering by flow

Read a snapshot of the data model (collections, fields, relations). Pass include_collections or exclude_collections for a partial snapshot.

Compare a schema snapshot against the current data model and report what applying it would change. Read-only — nothing is modified.

Apply a schema diff produced by diff_schema. Changes the live data model and can drop collections and fields.

Import rows from a CSV or JSON file into one collection, or into several related collections at once using a flat multi-collection file. Subject to the Directus IMPORT_MAX_FILE_SIZE limit (50mb by default).

Single-collection: pass collection; the file is a plain array of rows. Batch (Directus 12.2+): omit collection; the file MUST be a JSON array of { "collection": string, "items": object[] } entries, e.g. [{"collection":"authors","items":[{"name":"Ann"}]},{"collection":"articles","items":[{"title":"Hi","author":1}]}]

Find the tools on this server that match a task description, and return their full definitions so they can be called directly. Use this first when unsure which tool to use.