Problem
List commands currently dump the full API object for every item. This wastes tokens when an agent is just trying to orient.
Examples:
datasets returns 8 fields per item: name, description, settings{delete_protected}, expand_json_depth, slug, regular_columns_count, last_written_at, created_at
boards list includes the full panels[] array — nested objects with query IDs, SLO IDs, positions, etc. This can be very large.
columns list returns all 7 fields including description, hidden, created_at, updated_at
markers list returns all 9 fields including timestamps
For a list, an agent typically only needs enough to identify the item and decide what to do next.
Fix
Default list schemas:
datasets: slug, name
boards: id, name, type
columns: id, key_name, type
markers: id, message, type
derived_columns: id, alias, expression
Full objects remain available via the get-* commands.
Optionally add a --fields flag for agents that need additional fields in bulk.
Reference
AXI Principle 2 — Minimal default schemas
Problem
List commands currently dump the full API object for every item. This wastes tokens when an agent is just trying to orient.
Examples:
datasetsreturns 8 fields per item:name,description,settings{delete_protected},expand_json_depth,slug,regular_columns_count,last_written_at,created_atboardslist includes the fullpanels[]array — nested objects with query IDs, SLO IDs, positions, etc. This can be very large.columnslist returns all 7 fields includingdescription,hidden,created_at,updated_atmarkerslist returns all 9 fields including timestampsFor a list, an agent typically only needs enough to identify the item and decide what to do next.
Fix
Default list schemas:
datasets:slug,nameboards:id,name,typecolumns:id,key_name,typemarkers:id,message,typederived_columns:id,alias,expressionFull objects remain available via the
get-*commands.Optionally add a
--fieldsflag for agents that need additional fields in bulk.Reference
AXI Principle 2 — Minimal default schemas