fb-marketplaceresearch-item

name: fb-marketplace:research-item description: Use when you need to research a product online to find its retail price, used market value, official description, specifications, and measurements. Invoke after identifying brand/model from images.

Research Item for FB Marketplace

Every accurate FB Marketplace listing in our workflow starts here. We never guess prices or descriptions - we research them.

Requirements

Before proceeding, you MUST:

1. Have a product identifier (brand name, model number, or detailed description)
2. Know the product category (electronics, clothing, tools, furniture, etc.)
3. Commit to checking multiple sources - single-source pricing always fails

Research Workflow

digraph research_flow {
    rankdir=TB;
    node [shape=box];

    start [label="Start with product identifier" shape=ellipse];
    search [label="1. Web search for exact product"];
    find_retail [label="2. Find retail listings\n(Amazon, manufacturer, retailers)"];
    find_used [label="3. Search used marketplaces\n(eBay sold, FB Marketplace, Craigslist)"];
    extract_desc [label="4. Extract official description"];
    extract_specs [label="5. Extract specifications/measurements"];
    estimate_price [label="6. Calculate price range"];
    output [label="Return research summary" shape=ellipse];

    start -> search;
    search -> find_retail;
    find_retail -> find_used;
    find_used -> extract_desc;
    extract_desc -> extract_specs;
    extract_specs -> estimate_price;
    estimate_price -> output;
}

Tool Strategy

Prefer WebSearch over WebFetch. WebSearch returns rich summaries directly and is fast/reliable. WebFetch often hangs or fails on retailer sites (Costco, Amazon, etc.).

• Use WebSearch: For all pricing, specifications, and product info
• Use WebFetch: Only for manufacturer spec sheets or PDFs that WebSearch can't summarize
• If WebFetch fails or hangs: Move on immediately - use WebSearch results instead

Run searches in parallel when they don't depend on each other.

Research Steps

1. Search for Product Info (Run in Parallel)

IMMEDIATELY run these WebSearch queries in parallel:

• "[Brand] [Model]" specifications price (primary search)
• [Brand] [Model] site:ebay.com sold (used market pricing)
• [Brand] [Model] review (features and real-world info)

2. Extract Pricing from Search Results

WebSearch results typically include prices directly. Look for:

• Retail prices from Amazon, Home Depot, Walmart, Costco in search snippets
• eBay sold prices in search results
• Do NOT use WebFetch on retailer sites - they block or hang

Record:

• Current retail price (or last known if discontinued)
• Used/sold price range from eBay results

3. Extract Product Description

From WebSearch results, extract:

• Official product name/title
• Key features (bullet points)
• Product category
• Target use case

Never write our own marketing copy when official descriptions exist.

4. Extract Specifications

Always look for these - missing specs make listings fail:

• Dimensions: Height, width, depth, weight
• Capacity: Volume, wattage, size ratings
• Materials: Construction materials
• Model variants: Color options, size options
• Included accessories: What comes in the box

5. Calculate Price Recommendation

We always use this pricing table - no exceptions:

You MUST adjust based on:

• Local market demand
• Seasonality
• Completeness (all parts/accessories)
• Age of product

Output Format

Every research summary MUST follow this exact format:

## Research Summary: [Product Name]

### Product Identification

- **Brand**: [Brand]
- **Model**: [Model Number]
- **Full Name**: [Official Product Name]

### Pricing

- **Retail Price**: $XX (source)
- **Used Market Range**: $XX - $XX
- **Recommended Listing Price**: $XX - $XX

### Description

[Official product description or synthesized description from research]

### Specifications

- Dimensions: [H x W x D]
- Weight: [Weight]
- [Other relevant specs]

### Key Features

- [Feature 1]
- [Feature 2]
- [Feature 3]

### Sources

- [URL 1]
- [URL 2]

Common Mistakes