What problem does it solve? Analysts working with dlt pipelines often need to understand unfamiliar datasets and turn business questions into charts without manually writing SQL and visualization code from scratch. This Skill connects to a pipeline, profiles its tables, and produces a validated chart plan that a notebook builder can consume. ## Core Features & Use Cases - Pipeline Discovery and Profiling: Connects to dlt pipelines via dlthub MCP tools or Python fallback, listing tables, schemas, row counts, anomalies, and PII flags. - Chart Planning: Translates a business question into a chart spec (type, axes, aggregation, time grain) with validated SQL and altair code, one chart per invocation. - Structured Artifact Output: Writes or appends to a dated analysis_plan.md file with connection info, profile summary, questions, data gaps, and chart blocks for downstream notebook generation. - Use Case: Ask "what's the revenue trend" against your orders pipeline and receive a confirmed monthly line-chart spec with SQL and altair code, ready for the build-notebook step. ## Quick Start Ask the assistant to explore your dlt pipeline and answer a question like "explore my data -- what's the revenue trend by month?"
