What problem does it solve? Software requirements arriving from upstream intake often lack clear objectives, scope, or acceptance criteria, causing downstream development agents to build on ambiguous foundations. This Skill acts as Agent 2 of 15 in a Databricks multi-agent SDLC pipeline, checking requirement completeness and emitting a validated requirement artifact with numbered clarification questions before any build work begins. ## Core Features & Use Cases - Completeness and Ambiguity Checking: Detects missing objectives, scope gaps, and under-specified acceptance criteria against retrieved requirement text, without inventing answers. - Contract-Validated Output: Returns JSON matching a strict output contract (status, clarification_questions, citations, tool_calls, tokens) that the LangGraph orchestrator validates before handoff. - Lakehouse Artifact Persistence: Writes validated_requirement.json to Unity Catalog Volumes and upserts trace rows into state.artifacts and state.agent_runs for audit. - Use Case: A Jira ticket enters the pipeline via Agent 1; this agent validates the requirement package, finds the acceptance criteria are missing, and returns status needs_clarification with three specific questions, pausing the pipeline at the human approval gate. ## Quick Start Ask the agent to validate the requirement package for ticket PROJ-123 from the upstream artifacts volume and return the validated requirement JSON with any clarification questions.
