specification

Expands a PRD capability area into a functional specification with atomic, testable requirements.

3|Updated Feb 18, 2026
One-click install
npx skills add https://github.com/gener8v/gener8v.claude-skills --skill specification-gener8v
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: specification
Source: https://github.com/gener8v/gener8v.claude-skills/tree/main/skills/specification
Command: npx skills add https://github.com/gener8v/gener8v.claude-skills --skill specification-gener8v

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve? Turning a high-level PRD capability area into detailed, testable requirements is slow and error-prone, and without a consistent format downstream work like constraints analysis, technical design, and ticket breakdown loses traceability. This Skill produces a structured, implementation-agnostic functional specification with unique requirement IDs that the rest of the pipeline can reference. ## Core Features & Use Cases - Atomic Requirement Generation: Decomposes PRD capability bullets into atomic, testable [PREFIX]-REQ-XXX requirements with priority and change tags. - Non-Functional Requirements: Captures measurable NFR targets (latency, availability, observability) each with a named verification method. - Living Artifact Management: Amends existing specifications in place with append-only IDs, marking withdrawn requirements instead of deleting them, and updates the change brief with Adds/Modifies/Withdraws deltas. - Use Case: After planning a support-search change, run the Skill on the Search & Retrieval capability area to produce .gener8v/specifications/search-and-retrieval.md with tagged requirements, edge cases, open questions, and an approval workflow. ## Quick Start Ask the AI to spec out the Search & Retrieval capability area for the support-search change using the specification skill.

Frequently Asked Questions about specification

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

FAQPage Schema
How do I write a functional specification from a PRD?▼

Invoke the Skill with a capability area name and change slug, for example 'specification Search & Retrieval for support-search'. It reads the PRD and change brief, then writes a specification with atomic requirements, NFRs, edge cases, and open questions to .gener8v/specifications/.

What makes a good software requirement in a specification?▼

Each requirement should be atomic (one behavior, split on 'and'), testable (no subjective words like 'quickly'), and behavior-focused rather than interface-focused. Requirements get unique [PREFIX]-REQ-XXX IDs so tickets, constraints, and code annotations can reference them unambiguously.

Can I re-run the specification skill on an existing specification?▼

Yes. Re-running for a later change amends the living specification in place: existing IDs are preserved, new requirements are numbered above the current maximum, and removed ones are marked Withdrawn rather than deleted. An amended Approved specification returns to Draft until re-approved.

What is the difference between an NFR and a constraint?▼

An NFR is a measurable target the system must achieve, like p95 latency under 800 ms, with a named verification method. A constraint is a boundary the system must operate within, like a mandated platform or regulation, and belongs to the Constraints skill instead.

Why does lint report a requirement prefix used by two specifications?▼

Two capability areas chose the same 2-4 letter prefix, breaking ID uniqueness. If nothing cites the newer IDs yet, change its prefix; once tickets or code annotations reference them, withdraw the IDs and re-issue them under a new prefix, recording the deltas in the change brief.