AgentSkillsCN

Affiliate Link Finder

每当有新补充剂被添加到数据库时,自动在 Amazon/iHerb 上搜索评分最高的相关产品,并生成相应的联盟营销链接建议。

SKILL.md
--- frontmatter
name: Affiliate Link Finder
description: Automatically searches for the highest-rated supplement products on Amazon/iHerb and generates affiliate link suggestions whenever a new supplement is added to the database.

Affiliate Link Finder Skill

Purpose

This skill automates the process of finding optimal product listings and generating affiliate links when supplements are added to the app database.

Trigger

This skill should be activated whenever:

  • A new supplement is being added to MockSupplementRepository
  • The user requests an affiliate link update for an existing supplement
  • Bulk supplement data is being imported

Workflow

Step 1: Extract Supplement Information

From the supplement being added, identify:

  • Primary name (e.g., "Omega-3", "Magnesium Glycinate")
  • Specific form/type (e.g., "EPA/DHA", "Glycinate", "Citrate")
  • Target dosage range (if relevant for search)

Step 2: Web Search Strategy

Perform targeted searches using the search_web tool:

Amazon Search

code
Query: "{supplement_name} {form} supplement ADHD {dosage}"
Example: "Omega-3 EPA DHA supplement ADHD 1000mg"

Look for:

  • Products with 4+ star ratings
  • High review counts (500+)
  • "Amazon's Choice" or "Best Seller" badges
  • Third-party tested/certified products
  • Brands mentioned in search results (Nordic Naturals, NOW Foods, Thorne, etc.)

iHerb Search (Alternative/Backup)

code
Query: "{supplement_name} {form} highest rated iHerb"

Look for:

  • iHerb verified reviews
  • Products with detailed supplement facts
  • Brand reputation indicators

Step 3: Extract Product Identifiers

For Amazon:

  • Extract ASIN (Amazon Standard Identification Number)
    • Found in URL: amazon.com/dp/{ASIN}
    • Example: B001U4OYGY
  • Extract product title for verification

For iHerb:

  • Extract Product ID
    • Found in URL: iherb.com/{product-id}
    • Example: 81332

Step 4: Generate Affiliate Links

Use the AffiliateService class to generate tagged links:

dart
// For Amazon
final affiliateService = AffiliateService();

// Single region
final usLink = affiliateService.getTaggedAffiliateLink(
  supplementId: 'B001U4OYGY',  // ASIN from Step 3
  region: UserRegion.us,
);

// All regions
final allLinks = affiliateService.getAllRegionLinks('B001U4OYGY');

Result format:

code
US:  https://www.amazon.com/dp/B001U4OYGY?tag=adhdsupplements-20
UK:  https://www.amazon.co.uk/dp/B001U4OYGY?tag=adhdsupplements-21
EU:  https://www.amazon.de/dp/B001U4OYGY?tag=adhdsupplements-22

For iHerb:

code
https://www.iherb.com/{product-id}?rcode=ADHDSUPP
Example: https://www.iherb.com/81332?rcode=ADHDSUPP

Step 5: Present Recommendation

Provide output in this format:

markdown
## Affiliate Link Suggestion: {Supplement Name}

### Recommended Product
**Title:** {Product Title from Amazon/iHerb}
**Rating:** {Star Rating} ({Review Count} reviews)
**ASIN/ID:** {Identifier}
**Price:** ${Price} (if available)

### Generated Affiliate Links
- **US Amazon:** {US Link}
- **UK Amazon:** {UK Link}
- **EU Amazon:** {EU Link}
- **iHerb:** {iHerb Link} (if applicable)

### Add to Mock Repository
```dart
Supplement(
  id: '{generate-uuid}',
  name: '{Supplement Name}',
  description: '{Brief description}',
  referralUrl: '{US Link}',  // Default to US
  benefits: [...],
  dosageInstruction: '{Dosage}',
  sideEffects: [...],
  focusLevel: {1-5},
)

Notes

  • {Any special considerations}
  • {Alternative products if primary is unavailable}
code

## Example Execution

**User Input:**
"Add L-Tyrosine supplement to the database"

**Agent Actions:**
1. Search: "L-Tyrosine supplement ADHD 500mg amazon"
2. Identify top product: "NOW Foods L-Tyrosine 500mg"
3. Extract ASIN: `B0013OQGO6`
4. Generate links using `AffiliateService`
5. Present recommendation (see format above)

## Edge Cases

### No Amazon Listing Found
- Fall back to iHerb search
- Search for general category: "{supplement_name} best brand reddit"
- Suggest manual research with specific brand recommendations

### Multiple Strong Candidates
- Present top 3 options ranked by:
  1. Rating × Review Count score
  2. Price per serving
  3. Third-party certification (USP, NSF, ConsumerLab)

### International Availability Issues
- Note if product unavailable in UK/EU
- Suggest region-specific alternatives
- Update affiliate link map accordingly

## Quality Checks

Before finalizing recommendation:
- [ ] ASIN/Product ID is valid
- [ ] Product matches supplement specification (name, form, dosage)
- [ ] Rating ≥ 4.0 stars
- [ ] Review count ≥ 100 (or note if lower)
- [ ] Affiliate tag is correctly formatted
- [ ] Link opens to correct product page

## Maintenance

This skill should be reviewed quarterly to:
- Update affiliate tag IDs if changed
- Verify top-rated products are still available
- Add new platforms (e.g., Fullscript, Wellevate)
- Refine search queries based on accuracy