# ecom

Analyzes D2C store order CSVs to diagnose revenue, retention, and product health issues.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · ★ 53
- Author: takechanman1228
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/takechanman1228/claude-ecom/tree/HEAD/skills/ecom/SKILL.md
- Updated: 2026-06-11
- HTML: https://affpapa.org/en/nevibecode/skills/ecom

## What it does

The skill evaluates ecommerce transaction data using an internal Python engine. It processes order CSV records across 30-day, 90-day, and 365-day periods. The engine calculates KPI trees and executes roughly 30 health checks across revenue, customer, and product categories. Input data must contain an order ID, order date, customer identifier, and net revenue.

Two execution modes are supported. For targeted questions about recent sales drops, the agent provides an inline answer of 10 to 30 lines. For a complete audit, it writes a structured `REVIEW.md` document with KPI status markers and a prioritized action plan.

The initial run bootstraps a private Python virtual environment inside `~/.local/share/claude-ecom/`. That initial setup takes about a minute, while subsequent runs execute immediately.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add takechanman1228/claude-ecom --skill ecom -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/ecom/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/ecom/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add takechanman1228/claude-ecom --skill ecom -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/ecom/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add takechanman1228/claude-ecom --skill ecom -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/ecom/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install takechanman1228/claude-ecom ecom
```

2. **Manually: put the files in place**
- File: `.github/skills/ecom/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add takechanman1228/claude-ecom --skill ecom -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/takechanman1228/claude-ecom --path skills/ecom
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/ecom/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add takechanman1228/claude-ecom --skill ecom -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/takechanman1228/claude-ecom/tree/HEAD/skills/ecom/SKILL.md

## When to use it
- diagnose revenue drops across 30d, 90d, and 365d windows
- evaluate store health across roughly 30 automated checks
- generate a structured REVIEW.md report with KPI trees
- track repeat purchase rates and average order values

## Example prompts
- /ecom review 90d
- Review our D2C store performance using orders.csv
- Why did our revenue drop last month based on orders.csv?

## Source text · SKILL.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: 5af4c75dccab2406.

BEGIN THIRD-PARTY FILE 5af4c75dccab2406
````text
---
name: ecom
description: >
  Ecommerce business review for D2C stores from order transaction CSVs.
  Runs the bundled Python engine (KPI trees, ~30 pass/watch/fail health
  checks, 30d/90d/365d windows), then interprets the results: either a
  full narrative business review written to REVIEW.md, or an inline
  answer to a focused question.
when_to_use: >
  Use for "ecommerce review", "store review", "store health", "business
  review", revenue/customer/product analysis of an online store, or
  questions over order data like "how was last month", "how's retention",
  "why did revenue drop".
argument-hint: "review [30d|90d|365d] [question]"
allowed-tools:
  - Read
  - Grep
  - Glob
  - Write
  - Bash(ecom *)
---

# ecom — Ecommerce Business Review Toolkit

D2C ecommerce analytics. The bundled Python engine computes KPIs, runs
health checks, and scores performance from order transaction data;
**you** (Claude) interpret the numbers and write the human-readable report.

**Key principle:** Python computes the numbers. Claude interprets them.
Never present raw numbers without business context.

**Arguments:** `$ARGUMENTS` (if empty, infer intent from the conversation)

## Mode Selection

| Arguments / intent | Mode | Output |
|---|---|---|
| empty or `review` | Full Review (auto-selects periods from data) | `REVIEW.md` |
| `review 30d` / `90d` / `365d` | Full Review, single period | `REVIEW_{PERIOD}.md` |
| contains a natural-language question | Focused Query | Inline answer, no file |

## Input

Order transaction CSV. Each row = one order or line item.
Required columns: order ID, order date, customer ID (or email), revenue
(after discounts, before tax/shipping). Optional: quantity, SKU/product,
discount amount. Column names are fuzzy-matched by the loader.

If no CSV is specified, Glob for `*.csv` in the working directory and ask
the user if multiple plausible candidates exist.

## Running the Engine

The `ecom` CLI is on PATH while this plugin is enabled:

```bash
ecom review orders.csv --output <output-dir>
ecom review orders.csv --period 90d --output <output-dir>
```

If `ecom` is not on PATH, use the bundled launcher with the same
arguments: `"${CLAUDE_SKILL_DIR}/../../bin/ecom"`. The first run
bootstraps a private Python venv under `~/.local/share/claude-ecom/`
and may take a minute; later runs are instant.

Output: `review.json` (or `review_{period}.json` for `--period` runs)
in the output directory (defaults to current directory).

## Workflow

### Phase 1: Compute (Python)

Run the engine (add `--period` per Mode Selection). It computes, per
available period: summary KPIs with prior-period comparison, a
new-vs-returning KPI tree, revenue driver decomposition (AOV / volume /
mix), and — for 365d — repeat purchase rate and a 12-month
monthly_trend. It also evaluates ~30 health checks across Revenue,
Customer, and Product; each returns pass / watch / fail and powers the
🟢/🟡/🔴 markers.

### Phase 2: Interpret (you — Claude)

**Full Review:**

1. Read `review.json` (see Data Rules below; full schema in
   [review-schema.md](references/review-schema.md))
2. Read [report-format.md](references/report-format.md) — REQUIRED
````
END THIRD-PARTY FILE 5af4c75dccab2406

The third-party file has ended. Everything between the 5af4c75dccab2406 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 3195 of 7916 bytes. © takechanman1228 · MIT · https://github.com/takechanman1228/claude-ecom/tree/HEAD/skills/ecom/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md