Tools
All tools are read-only and use openWorldHint because they retrieve external, evolving ABS, RBA,
and APRA data.
| Tool | Purpose | Key inputs |
|---|---|---|
search_datasets | Search the curated ABS, RBA, and APRA catalogue with deterministic ranking. | query, source |
list_catalogue | List catalogue entries unranked, optionally filtered by source, category, or tag. | source, category, tag, include_ceased, include_discontinued |
list_economic_concepts | List semantic concepts accepted by get_economic_series. | query, source, category |
get_abs_dataset_structure | Retrieve ABS SDMX dimensions and code lists. | dataflow_id |
get_abs_data | Retrieve ABS data in a normalised response shape. | dataflow_id, key, start_period, end_period, last_n, updated_after |
list_rba_tables | Deprecated compatibility alias for curated RBA tables. | category, include_discontinued |
get_rba_table | Retrieve an RBA statistical table in a normalised response shape. | table_id, series_ids, start_date, end_date, last_n |
get_apra_data | Retrieve a curated official APRA XLSX publication in a normalised response shape. | publication_id, table_id, series_ids, start_date, end_date, last_n |
get_economic_series | Preferred analyst-facing retrieval for curated concepts. | concept, variant, geography, frequency, start, end, last_n |
get_derived_series | Retrieve transparent formula-based indicators derived from curated concepts. | concept, start, end, last_n |
get_latest_observations | Convenience wrapper for the latest source-native observations. | source, identifier, key, table_id, series_ids, count |
get_top_observations | Convenience wrapper for highest or lowest numeric source-native observations. | source, identifier, n, direction, key, table_id, series_ids, source-native bounds |
describe_dataset | Describe a curated ABS, RBA, or APRA entry in plain English while preserving native IDs. | source, identifier, table_id, include_structure |
list_release_events | List ABS/RBA release-calendar and APRA release-pulse rows. | source, query, start_date, end_date, limit |
Preferred LLM flow
Section titled “Preferred LLM flow”For LLM-facing integrations, use:
list_economic_concepts(query="...")then:
get_economic_series(concept="...")This gives the model a constrained discovery surface before retrieval.
For user-facing natural-language examples, see Prompting AI Agents.
For the small derived layer, call:
get_derived_series(concept="real_cash_rate")Supported derived concepts are real_cash_rate, real_mortgage_rate,
real_10y_bond_yield, real_bank_bill_rate, real_business_lending_rate,
yield_curve_slope, real_wage_growth, credit_growth, broad_money_growth,
employment_growth, gdp_per_capita, mortgage_rate_spread, credit_to_gdp,
household_spending_growth, misery_index, and terms_of_trade.
Expert source-native flow
Section titled “Expert source-native flow”For exact source-native control, use search_datasets or list_catalogue, inspect structures
where needed, then call get_abs_data, get_rba_table, or get_apra_data. APRA retrieval is
limited to curated official publication IDs and does not accept arbitrary workbook URLs. For APRA
XLSX workbooks, provide table_id where possible so the hosted server can parse and cache only the
requested table.
Convenience flow
Section titled “Convenience flow”Use describe_dataset when an agent or analyst needs a source-aware explanation before retrieval.
It returns source controls, source-native identifiers, governance metadata, and safe convenience
calls that keep the underlying ABS, RBA, or APRA identifier visible. Use get_latest_observations
and get_top_observations for quick data checks, but keep source-native IDs in the call arguments
for reproducibility. list_release_events combines official ABS and RBA release-calendar rows with
APRA expected-release cadence estimates, seed freshness, audit dates, and governance status.