What problem does it solve? Teams often jump straight into implementation without clarifying requirements, comparing alternatives, or recording decisions, leading to rework and undocumented trade-offs. This Skill enforces a spec-first delivery lifecycle so every engineering, AI, FDE, Databricks, or content change begins with structured discovery and approved design artifacts. ## Core Features & Use Cases - Structured Discovery: Requires 5–10 concise discovery questions, then restates answers, assumptions, scope, non-goals, and success measures for approval. - Mandatory Design Artifacts: Produces a specification, an RFC comparing alternatives, an editable .excalidraw architecture diagram, and an ADR recording the approved decision before any implementation. - Gated Incremental Delivery: Breaks work into tracer-bullet increments with tests and evidence, blocking implementation until all five gates are approved. - Use Case: When asked to add a new Databricks pipeline, the Skill first gathers discovery answers, drafts the spec/RFC/ADR set, and only then permits coding with recorded test evidence. ## Quick Start Ask the agent to plan and deliver a new feature using the spec-first delivery workflow with discovery questions, RFC, architecture diagram, and ADR before implementation.
