cs-req

Writes and updates plain-language requirement documents describing existing product capabilities.

Updated Jun 8, 2026
One-click install
npx skills add https://github.com/wildlily1021/XW_Software --skill cs-req-wildlily1021
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cs-req
Source: https://github.com/wildlily1021/XW_Software/tree/main/.agents/skills/cs-req
Command: npx skills add https://github.com/wildlily1021/XW_Software --skill cs-req-wildlily1021

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams often lack clear, non-technical records of why each existing capability exists, what problem it solves, and where its boundaries lie. This Skill maintains a codestable/requirements/ directory of user-story-based requirement documents that describe current capabilities in plain language, keeping them synchronized with the actual implementation. ## Core Features & Use Cases - Backfill mode: Creates a requirement document for a capability that already exists in code but was never documented, using user stories, pain points, solution overview, and boundaries. - Update mode: Refreshes an existing requirement document when the capability's user story, pitch, or boundaries have evolved, with a changelog entry for structural changes. - Quality self-check: Enforces a review checklist that rejects PRD-style field dumps, implementation details, vague user stories, and missing boundaries before user review. - Use Case: After a feature acceptance phase adds a new user-facing capability, trigger backfill to produce a one-page requirement doc with a pitch usable as marketing copy, linked to the architecture doc via implemented_by. ## Quick Start Ask the assistant to backfill a requirement document for the data monitoring capability that was just accepted, describing why it exists and its boundaries in plain language.

Frequently Asked Questions about cs-req

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I document requirements for an existing feature?▼

Use backfill mode to create one requirement document per capability. It reads user materials, existing requirement docs, and related architecture docs, then drafts user stories, pain points, solution overview, and boundaries in plain language for your review.

What is the difference between a requirement doc and an architecture doc?▼

Requirement docs explain why a capability exists and what problem it solves in non-technical language. Architecture docs explain how it is built technically. The requirement doc links to its architecture doc through the implemented_by frontmatter field.

When should I not write a requirement document?▼

Skip requirement docs for pure internal refactors, tech debt cleanup, or tooling changes that add no user-facing capability. Also do not document planned features; those belong in a roadmap until the capability actually exists.

Can I update multiple requirement documents at once?▼

No, the workflow enforces a single-document rule per run. Processing one document at a time ensures each gets a thorough user review instead of batch-merging shallow drafts that nobody reads carefully.

Why does my requirement draft get rejected in the self-check?▼

Common rejection reasons include PRD-style field stacking, lecture-like tone, metaphorical titles, abstract user stories, embedded implementation details, and missing boundaries. Each issue must be fixed by rewriting, deleting, or moving content before user review.