dr-anomalies

Datarails Anomaly Detection

Find data quality issues in Finance OS tables. The MCP server's profile tools are deliberately thin — they return baseline aggregates and nothing more. This skill computes the actual findings (outliers, severity, duplicate counts, null rates) by post-processing those aggregates. Be explicit about which numbers came from the tool versus which the skill derived.

Tool reality check

Before designing the analysis, read what each tool actually returns:

Tool	What it returns	What it does NOT compute
profile_numeric_fields	SUM, AVG, MIN, MAX, COUNT per numeric field — in the backend-native DR_Values/col_keys/row_keys layout, no per-value aggregator labels	median, std dev, percentiles, outlier flags, null counts
profile_categorical_fields	distinct-count + first 10 sample values per field (capped at 5 fields; bare calls default to upload/mapping metadata columns — always pass fields)	per-value frequency, null counts, uniqueness ratio
start_aggregation_by_id / …_by_alias → poll get_aggregation_result_by_id / …_by_alias	grouped totals with no row limit	anything not expressible as GROUP BY + aggregation
get_data_by_id / …_by_alias	raw rows (≤500/page) with value-list and advanced filters	—

For severity, percentiles, std dev, z-scores, and most named anomaly categories below — the skill computes them, not the tool. There is no server-side anomaly tool; everything is derived here.

Workflow

Step 1: Verify Authentication

If any tool call fails with an authentication or connection error, guide the user to connect via the Connectors UI ("+" → Connectors → Datarails → Connect).

Step 2: Gather baseline aggregates

1. get_fields_by_id(table_id) — field ids, names, types. (If you only have a name/alias, resolve the table via list_data_models first.)

Async fetch — aggregations and distinct values run as start → poll. start_aggregation_by_id/_by_alias and start_distinct_values_by_id/_by_alias take the same arguments as the retired blocking calls (dimensions/metrics/filters; table id + field id, or alias + field alias) and return immediately with {"status": "pending", "handle": {...}}. Echo that handle back verbatim to the matching get_aggregation_result_by_* / get_distinct_values_result_by_* tool: a {"status": "running", "retry_after_seconds": N} response means poll again with the same handle after ~N seconds (≈5s) — it is not an error, and large jobs may take several polls; when ready, the result arrives in the familiar shape (for distinct values, pass limit to the result tool). An expired/unknown-handle error means restart with the start_* tool. Transitional fallback: if the start_* tools aren't available on the connector (older server), the blocking twins get_aggregated_data_by_* / get_distinct_values_by_* still work with the same arguments.