What problem does it solve?
Writing a PRD and its test cases separately often produces misaligned documents: requirements that cannot be tested, test cases that invent fields not present in the code, and review HTML that silently drops content. This Skill co-produces a story-driven PRD and atomic, code-grounded test cases as one workflow, keeping both documents strictly consistent.
Core Features & Use Cases
- Dual-document generation: Produces four artifacts per requirement — a PRD markdown and a test-case markdown (fact sources for AI) plus two template-based review HTML files for humans, verified 1:1 against the markdown via a scripted check.
- Code-grounded test cases: Every assertion cites a
file:line code reference, each case is a single atomic verification point with 13 fields, and task-type cases specify exact per-round prompts and expected results.
- Structured collaboration flow: A six-phase loop (requirement confirmation, autonomous code reading, PRD story discussion, test case drafting, adversarial review by independent sub-agents, freeze and registry) with single-point user confirmation at each step.
- Use Case: After a feature task is scoped, ask the agent to write the PRD and test cases; it reads the codebase first, drafts user stories with ASCII wireframes and Mermaid diagrams, then produces executable test cases with evidence types and failure classification.
Quick Start
Ask the agent to write the PRD and executable test cases for your already-scoped task, and confirm each stage as it walks through the six-phase workflow.
