shah-mohammed/ecommerce-claude-skills

inventory-risk-brief-ecommerce

Turns e-commerce inventory and sales exports into a practical risk brief for stockouts, overstock, promo conflicts, and traffic allocation.

查看源码

使用方法

  1. 复制安装命令,在自己的智能体环境安装前先检查源码包。
  2. 从原始文档核对所需工具、凭据和支持的宿主。
  3. 说明输入材料与期望交付物,并对照来源检查输出。
npx skills add https://github.com/shah-mohammed/ecommerce-claude-skills --skill inventory-risk-brief-ecommerce

开始前需了解

  • 安装不包含模型订阅、API 额度或访问文件的权限。
  • 仓库星数和安装量反映活跃情况,不等于独立质量或安全测试。
仓库原始内容

按源仓库内容呈现,保留标题、案例、代码、表格、链接以及原文引用的演示图片。

Inventory Risk Brief - E-commerce

Use this skill when

The user wants an inventory-aware growth or promo review from exports.

Common requests:

  • "Review this inventory export."
  • "Which products are at risk of stockout?"
  • "Which products need promo support?"
  • "What should we avoid pushing in ads?"

Required input

Minimum useful input:

  • Inventory export with SKU, product name, current stock, and variant.
  • Sales export for the last 30, 60, or 90 days.
  • Date range and sales velocity.
  • Promo calendar or planned campaign notes if available.

Recommended columns:

  • SKU
  • product name
  • variant
  • current inventory
  • units sold
  • revenue
  • margin or priority flag if available
  • incoming stock quantity and ETA if available
  • supplier lead time if available

Before analysis

  1. Confirm the purpose: prevent stockout, reduce overstock, plan promo, or protect paid traffic.
  2. Ask for exports if the user only describes inventory from memory.
  3. Mark missing lead time, incoming stock, and margin data as caveats.
  4. Do not pretend to run demand forecasting.

Analysis workflow

  1. Normalize SKU and product names.
  2. Calculate simple signals where data allows:
  • recent sales velocity
  • days of inventory remaining
  • sell-through risk
  • slow mover status
  • promo conflict risk
  1. Classify products:
  • stockout risk
  • overstock risk
  • safe to promote
  • do not scale traffic
  • needs more data
  1. Tie risks to marketing action.
  2. Name what cannot be decided yet. Reorder quantity needs lead time and incoming stock; a promo decision needs margin. Where those are absent, hand the owner the question rather than a recommendation built on a guess.

Decision and evidence standard

Every finding should include:

  • Evidence tag: export, screenshot, url, review_cluster, support_ticket, return_reason, policy, feed_diagnostics, margin_csv, inventory_export, hypothesis, or needs_data.
  • Severity: low, medium, high, or critical.
  • Confidence: low, medium, or high.
  • Business impact: revenue, margin, cashflow, retention, conversion, support_load, or risk.
  • Effort: XS, S, M, or L.
  • Owner decision: donow, test, investigate, monitor, ignore, or approvalneeded.

If the evidence is weak, mark the finding as hypothesis or needs_data and lower confidence.

Output format

Inventory verdict

Short owner-level summary.

Risk table

Product/SKUStatusEvidenceMarketing implicationRecommended action

Promo notes

Products suitable for promotion, products to protect, and products to avoid scaling.

Missing data

Fields needed for better confidence.

Example input and output

Input:

  • inventory export
  • 60-day sales export
  • incoming stock ETA
  • promo calendar

Good output excerpt:

FindingEvidenceSeverityConfidenceBusiness impactEffortOwner decision
Bestseller SKU may stock out during planned campaigninventory_export, exporthighmediumrevenue/riskSapproval_needed
Slow-moving variant has high stock and low recent sales velocityinventory_export, exportmediummediumcashflowStest

What not to do yet: infer stock risk without inventory or sales exports, or recommend purchase orders without lead time and approval.

Guardrails

  • Requires an inventory or sales export. Do not infer stock risk from product names alone.
  • Do not replace ERP, inventory planning, or demand forecasting systems.
  • Do not recommend purchase orders without lead time, margin, and business approval.
  • Do not recommend pushing products that may stock out during the campaign.
  • Do not ignore variant-level constraints.