daana-code/daana

Use when answering a business analytics question from approved reporting outputs, including totals, slices, distinct counts, or effective-time and knowledge-time comparisons, without changing the model.

Use when binding reviewed DAB meaning or a labelled source-led candidate to source evidence, authoring mapping YAML, or changing source strategy and effective-time expressions.

Use when authoring DAB business entities, attributes, relationships, identity or time definitions, or drafting a first candidate from typed sources or legacy dbt/Dataform code without a business owner present.

Use when translating an approved analytical design into a Daana DAR specification or changing an existing declaration without changing its reviewed business meaning.

Use when adding or changing a DAS data contract, field description, source path, type, key, model, or materialization declaration.

Use when implementing or changing a dlt landing adapter, provider acquisition boundary, observation envelope, or landing regression tests.

Use when an analytics pipeline fails or layers disagree and the task is to locate the owning boundary, explain the evidence, and prepare a safe next step rather than perform recovery.

Use when a reviewed business question needs a DAR fact grain, dimensions, temporal interpretation, reuse decision or runtime metric design, before choosing compiler output kinds.

Use when interviewing a business owner, clarifying an analytics decision or question, reviewing business definitions conversationally, or facilitating a question-led Agile Analytics loop. Technical model/mapping implementation uses the layer skills after meaning is reviewed.

Use when tracing DAB physical rows, hidden duplicates, effective/version history or unexpected processing windows to the producing pipeline and workflow using metadata.

Use when changing DAS generator behavior, source-shape mechanics, contract validation rules, or generated-output regression tests.

Use when preparing or performing warehouse inspection, SQL analysis, profiling, or source-to-output reconciliation across data-platform layers.

Use when preparing, executing, refreshing, or recovering a Daana DAB workflow or deployment, including scoped operations and lineage investigation.

Use when preparing, executing, retrying or reconciling a Daana DAR warehouse operation after analytical meaning is approved. Local-only check, lint and compile use structural validation instead.

Use when preparing, executing, retrying, or verifying an explicitly scoped Dataform materialization or rebuild.

Use when preparing, running, or recovering a dlt landing load, historical backfill, or persisted pending load package.

Use when checking DAR declarations, local generated SQL or dependency/schema readiness, including cases where a successful compile may conceal missing deployed columns or invalid published object shape.

Use when investigating DAS mapped values, source-present NULL loss, keys, latest winners, parity, or deployed table and partition behavior.

Use when checking DAS contracts, generated drift, Dataform compilation, or BigQuery SQL analysis before a PR or operation.

Use when checking DAB source-to-base-to-serving results, timelines, flow totals, relationship cardinality, duplicates, or post-operation reconciliation.

Use when independently reconciling a DAR result, grain, measures, temporal joins or dimension coverage against approved requirements and source evidence, including synthetic and deployed comparisons.