# Data Acquisition Design

> Use when the user needs to decide what data to collect before scraping or API work: DatasetNeed, DatasetSpec, entity grain, required vs nice-to-have fields, freshness, history, coverage targets, join keys, exclusions, and uselessness criteria. Use for vague business goals, all data requests, and scope control before source discovery.

- Skill: `pranjay-kumar/data-acquisition-design` (Agent Skill)
- Install (CLI): `npx skillmds@latest add pranjay-kumar/data-acquisition-design`
- Raw SKILL.md: https://api.skillmd.com/api/skills/pranjay-kumar/data-acquisition-design/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: Pranjay-kumar (https://skillmd.com/u/pranjay-kumar)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/pranjay-kumar/data-acquisition-design

---


# Data Acquisition Design

Act as the dataset designer for acquisition work. Convert vague goals into the smallest useful dataset.

## Shared Core

Read from `../data-acquisition-core/references/`:

- `source-access.md`
- `output-contracts.md`
- `workflow.md`
- `examples.md`

## Output

Return:

- `ModeSelection`
- `SourceAccessClass` when access is already implied
- `DatasetNeed`
- `DatasetSpec`
- `DataAcquisitionMemo`
- `ApprovalGate`

Do not hunt endpoints unless the user asks to continue.