ecomrads-virality-analysislisted
Install: claude install-skill Claudeforecommerce/claude-for-ecommerce
# Virality Analysis
Analyze a finished ad and predict its performance with the eComrads **`analyze_ad`** tool (`POST /post/ad-score`). This is **analysis in → text report out** — not a generated media asset.
> Naming: the customer-facing feature is **Virality Analysis**. The underlying MCP tool is still named **`analyze_ad`** (route `/post/ad-score`) — call that, but present results as "Virality Analysis."
Read [./AGENTS.md](./AGENTS.md) first.
## Workflow
1. **Upload** the ad creative → `upload` → public link (or use a link the user gives). Sandbox paths don't work.
2. Call **`analyze_ad`** with the link.
3. **Poll `check_generation`** until terminal.
4. **Deliver the summary** in plain language as a Virality Analysis.
## Call
Fields map to `AdScoreRequest`:
```json
{
"body": {
"media_url": "<url from upload>",
"media_type": "image",
"context": "DTC skincare, target 25–34 women, IG feed"
}
}
```
- **`media_url`** (required) — the creative to analyze (link from `upload`). This tool takes **only** a URL — never base64 or a chat image directly.
- **`media_type`** — `image` (default) or `video`.
- **`context`** — optional: audience, platform, offer, goal. Improves the analysis.
## Delivering
Report the result from the **completed** job as a short, plain-language Virality Analysis — headline score, the strongest elements, the biggest weaknesses, and 1–3 concrete fixes. Never fabricate numbers; use only what the job returns.
```
Virality Analysis: 72/10