quantumbyteoss/quantumbyte/business-requirements

Claude Code and agent skill by quantumbyteoss/quantumbyte

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Read and hand-edit the app's business requirements — the checklist the harness verifies the built app against. MUST be loaded whenever the user asks to add, change, remove, reword, re-prioritise, or reclassify a business requirement (e.g. "add a requirement that…", "that rule is wrong, fix it", "drop the requirement about X", "make the closed-outlet rule critical"). The requirements are a plain JSON file in the app repo; edit it directly with the file tools, then create the lock marker so the c…

Install

npx vskill@latest install quantumbyteoss/quantumbyte/business-requirements

SKILL.md

Business Requirements

The harness checks the finished app against a list of business requirements and gives a ship / do-not-ship decision. Those requirements live in the app repo as a single JSON file:

_doc/business-requirements.json

It is a JSON array; each entry is one requirement:

json
{
"id": "no-checkout-when-closed",
"excerpt": "No buying from a closed store.",
"description": "A customer cannot check out when the selected outlet is closed.",
"type": "explicit",
"severity": "critical"
}

Field rules:

  • `id` — stable kebab-case slug. Reuse the existing id when editing a requirement (never

renumber). Only mint a new kebab-case id for a genuinely new requirement.

  • `excerpt` — a tiny glance-line, ideally under ~8 words, clearly shorter than description.
  • `description` — the requirement in explain-like-I'm-five language: short sentences,

everyday words, as if explaining to someone with zero technical background. Describe what a person or the business experiences — never code, functions, APIs, endpoints, databases, schemas, tables, or frameworks. If you catch yourself naming a technical thing, rephrase it as what the user or business experiences.

  • Good: "A customer cannot check out when the selected outlet is closed."
  • Good: "Loyalty points are only awarded after an order is completed, never when it is cancelled."
  • Bad: "The /checkout endpoint validates outlet.is_open before creating the order."
  • `type` — "explicit" when the user asked for it directly, "hidden" when it is implied by

the product category but not stated (e.g. preventing double rewards, blocking orders to a closed outlet).

  • `severity` — "critical", "high", "medium", or "low", by business impact if unmet.

Only critical requirements block the ship gate; the rest surface as warnings.

How to edit (MANDATORY steps)

Requirements are normally auto-derived from _doc/product-overview.md. To make a hand edit stick — so the next auto-derivation does not overwrite it — you MUST do BOTH:

  1. Edit the file. Read _doc/business-requirements.json, apply the add / change / remove

the user asked for (keep it a valid JSON array, preserve unrelated entries and their ids), and write it back.

  1. Write the lock marker. Create/refresh the file

_doc/business-requirements.lock (any short text content). Its mere presence tells the system the requirements were hand-edited so auto-derivation stops overwriting them. Skipping this means the user's edit will be silently regenerated away.

Both files live under _doc/, so the end-of-turn commit persists them. After the turn, the harness automatically re-checks the app against the new requirement set — you do not trigger it.

When NOT to touch this

  • If the user is changing what the app does (a feature, a rule in the product), fix the app —

that is normal build work, and the harness re-checks it afterward. Only edit business-requirements.json when the user is changing the requirement text itself (the thing being checked), not the implementation.

  • Never invent requirements the user didn't ask for; never delete requirements they didn't

mention.

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/quantumbyteoss/quantumbyte/business-requirements/badge)](https://verified-skill.com/skills/quantumbyteoss/quantumbyte/business-requirements)

Something wrong with this skill or its page? Tell us.