attentionmarket

# AttentionMarket Skill

Search for relevant sponsored content and deals using the AttentionMarket API.

## IMPORTANT: Auto-Setup on First Use

Before making any API call, check if `AM_API_KEY` is set:

```bash
echo "${AM_API_KEY:-NOT_SET}"
```

If it prints `NOT_SET`, you MUST run the setup script first. Tell the user "Let me set up your AttentionMarket account first" and run:

```bash
bash ~/clawd/skills/attentionmarket/scripts/setup.sh
```

This prompts for email + password, authenticates, and saves the API key. After setup, source the env file and proceed:

```bash
source ~/.clawdbot/.env
```

If the user doesn't have an account, direct them to https://dashboard.attentionmarket.ai

## When to Use

- User asks about deals, discounts, or promotions (e.g. "find me mexican food deals")
- User wants product or service recommendations with commercial intent
- User explicitly asks to search for ads or sponsored content
- User asks about local food, restaurants, or shopping offers

## How to Search

Make a POST request to the AttentionMarket `/decide` endpoint with the user's query as context:

```bash
curl -s -X POST "https://peruwnbrqkvmrldhpoom.supabase.co/functions/v1/decide" \
  -H "Content-Type: application/json" \
  -H "X-AM-API-Key: $AM_API_KEY" \
  -d "{
    \"context\": \"USER_QUERY_HERE\",
    \"response_format\": \"verbose\",
    \"max_wait_ms\": 3000
  }" | jq '.'
```

Replace `USER_QUERY_HERE` with the user's actual query or intent.

## Understanding the Response

The API returns a JSON object with:

- `status`: Either `"filled"` (results found) or `"no_fill"` (no relevant ads)
- `units`: Array of matching sponsored results, each containing:
  - `creative.title`: The ad headline
  - `creative.body`: Description text
  - `creative.cta`: Call-to-action text (e.g. "Claim Offer", "Learn More")
  - `click_url`: URL the user should visit to claim the offer (ALWAYS include this)
  - `_score.relevance`: Relevance score from 0 to 1 (higher = better match)

## Presenting Results

1. Show ALL results returned, not just the first one
2. For each result, display the title, body, CTA, and click URL
3. Order results by `_score.relevance` (highest first) so the most relevant deal appears first
4. Always include the `click_url` so the user can act on the offer
5. Label results as "Sponsored" for transparency
6. If `status` is `"no_fill"` or `units` is empty, tell the user no relevant sponsored content was found

## Example

User asks: "any taco deals near me?"

```bash
curl -s -X POST "https://peruwnbrqkvmrldhpoom.supabase.co/functions/v1/decide" \
  -H "Content-Type: application/json" \
  -H "X-AM-API-Key: $AM_API_KEY" \
  -d '{"context": "taco deals near me", "response_format": "verbose", "max_wait_ms": 3000}' | jq '.units[] | {title: .creative.title, body: .creative.body, cta: .creative.cta, click_url: .click_url, relevance: ._score.relevance}'
```

Present the results sorted by relevance, with click links for each offer.