Gathering Architecture Requirements

作者 gopigment6fec49f4ce9d無授權條款22 個星標收錄於 2026年10月8日更新於 2026年10月8日儲存庫2 天前更新

Planning skill. Use before building anything to gather requirements through structured discovery questions. Covers dimensions, UX, data, governance, and planning cycles. Load this skill during Phase 0 of building-a-full-application.

AI 產生的概覽

透過結構化探索提問,在打造規劃應用前蒐集架構需求。

功能
這項規劃技能會依序針對維度與結構、終端使用者 UX、資料來源與週期、治理與安全,以及規劃週期與版本等面向,提出結構化的探索提問。最後提供一份驗證清單,用來在定案架構前確認維度結構、資料粒度、治理與版本管理。它產出的是蒐集到的需求與決策,而非檔案或程式碼。
適用情境
在打造任何東西之前、規劃或業務應用專案的需求蒐集階段使用。它設計用於打造完整應用的 Phase 0,在撰寫規格之前載入。
執行需求
不需要指令碼或工具,僅為指示性內容。它假定處於代理環境中,並引用一個規劃平台及一項相關的應用打造技能。

Gathering Architecture Requirements

Work through each area systematically, gathering requirements and making decisions before proceeding to the spec.


Dimensions and Structure

  • At what granularity will users input data for each business dimension?
  • At what granularity will calculations occur?
  • Do dimension item mappings change over time (employee moving cost centers)?
  • What level drives user access rights?
  • Are there organizational hierarchies with varying granularity needs?
  • Which dimensions are required in the structure? Can any be replaced with properties or mapped dimensions?

End User UX

  • Can you provide wireframes or mockups of desired input screens?
  • Can you share examples of current reports before Pigment?
  • On which dimensions do users need to input?
  • What are the key final reports (P&L, etc.)?
  • What variations need to appear as columns in reports?
  • What KPIs or ratios are required, and how are they calculated?
  • Do certain calculations require separate metrics vs calculated items?
  • What workflow steps exist, and do access rights need to change between steps?

Data Sources and Cycle

  • What are all data sources (ERPs, HR systems, CRM, etc.)?
  • At what granularity is data available for each dimension?
  • At what granularity do users need to plan?
  • What is the gap between available data and planning granularity?
  • Where will data transformation occur (before or within Pigment)?
  • What metadata is available for each dimension?
  • What historical data needs to be loaded, and is it transactional or aggregated?

Governance and Security

  • Which teams will manage which use cases?
  • What data is sensitive and requires restricted access?
  • Do any use cases involve individual-level data?
  • Which dimensions are shared across multiple use cases?
  • Are any dimensions created by business logic within a specific use case?
  • Which users need access to which data?
  • Do users need to navigate between applications?

Planning Cycles and Versions

  • What versions/scenarios are needed?
  • How many live versions concurrently (max ~10 recommended)?
  • How should completed plans be protected from changes?
  • Do users need to replan based on historical budgets?
  • When loading new data, should it impact historical plans?
  • What variations need to appear as columns in reports?
  • How will new planning cycles be initialized?
  • Do transaction lists need effective dating?

Validation Checklist

Before finalizing architecture, validate:

  • Can the dimensional structure support all required inputs and calculations?
  • Does the structure enable the desired UX?
  • Does data granularity match structural requirements?
  • Are governance and security requirements met?
  • Is version management adequate for the planning cycle?
  • Are there fewer than ~10 live versions?
  • Have all dimensions been challenged for necessity?

來源與署名

來源:gopigment/ai-plugins位於skills/gathering-architecture-requirements提交6fec49f

授權條款: 無授權條款

內容歸原作者所有。SourceWeft 從公開儲存庫中收錄這些內容。

檢舉或申請下架