What problem does it solve?
Raw stakeholder input is noisy, fragmented, and often mixes functional asks with implementation hints, making it hard for architects and product teams to produce a clear, reviewable set of requirements; this skill standardizes and disambiguates that input into traceable requirement documents ready for design.
Core Features & Use Cases
- Interactive extraction: Reads PDFs, DOCX, markdown notes, emails, and call transcripts to surface atomic candidate requirements with source quotes.
- Structured output & parking: Proposes domain/feature breakdowns, writes frontmatter-compliant Markdown files under docs/requirements/<domain>/<feature>.md, and copies source documents into per-domain sources folders while preserving originals.
- Traceability & guardrails: Deduplicates candidates, flags open points and disagreements, prevents silent domain creation, and appends ingestion history for incremental updates.
- Use case: Hand a collection of kickoff briefs and a meeting transcript to the skill to produce a validated set of draft requirement files that an architect can immediately consume.
Quick Start
Ingest these files and convert them into docs/requirements entries organized by domain and feature: provide the paths to the stakeholder brief and call transcript and confirm the proposed breakdown.