name

amazon-listing-cosmo

description

Amazon US listing research and execution workflow for competitor ASIN audits, COSMO-style scenario intent mapping, FABE benefit rewriting, keyword scoring, optimized titles/bullets, and 7-image listing plans or image prompts. Use when the user asks to audit Amazon competitors, build or optimize an Amazon listing, analyze ASINs, produce keyword score tables, write title/bullets, generate listing image briefs, or create Amazon listing assets from product data.

Amazon Listing COSMO

Core Rule

Do not invent Amazon data, product specs, certifications, or performance claims. If a field cannot be verified from provided product data, fetched pages, or user-supplied files, mark it as 未获取 or remove it from publish-ready copy.

Never use competitor brand names in final title, bullets, backend terms, or image text. Use competitor names only in the audit tables when they are source data.

Workflow

Gather inputs
- Product category, target marketplace, target product specs, brand restrictions, and competitor ASINs.
- If ASINs or URLs are provided, browse/fetch current pages. Amazon content changes, so do not rely only on memory.
- If the user asks for image generation, request or use verified product reference data. If unavailable, use only specs explicitly verified in the competitor/product source and disclose the limitation.
Competitor audit
- Extract per ASIN: title, bullets, product details, A+ selling points, image sequence logic, review pain points, and high-frequency user scenarios.
- Use 未获取 for missing fields.
- Produce these tables: competitor basic info, selling point matrix, image expression matrix, review pain point matrix, borrowable points, avoid points.
- For the detailed table pattern, see references/audit-template.md.
COSMO intent alignment
- Build semantic shopping intent relationships, not just keyword stuffing:
  - Product is_a
  - Used_for_func
  - Used_for_audience
  - Used_in_location
  - Used_on_body/object
  - Capable_of
  - Used_with
- Use these relationships to connect product, problem, scene, audience, and accessory context.
- For the scoring model, see references/cosmo-fabe-keywords.md.
FABE benefit rewriting
- Convert parameters into user-perceived benefits:
  - Feature: what it is
  - Advantage: why it is better or easier
  - Benefit: what the buyer feels or accomplishes
  - Evidence: what image/detail/proof supports it
- Avoid medical, absolute, unsupported, or safety claims.
High-quality listing output
- Create keyword value score table.
- Create Amazon title with highest-value terms in the first 80 characters.
- Write bullets in this structure: CORE BENEFIT IN CAPS - scenario-based detail.
- Provide English and Chinese bullets if the user asks in Chinese or requests bilingual output.
- Include backend search terms and compliance exclusions.
- For title/bullet patterns, see references/listing-copy-template.md.
7-image listing plan
- Build exactly this sequence unless the user requests another:
  1. Main Image
  2. Lifestyle Image
  3. Feature Infographic
  4. Pain Point Solution
  5. Compatibility / Size
  6. Material / Detail
  7. Use Case / Final Conversion
- For each image, output: 图序, 图片类型, 转化目标, 对应用户痛点, 核心关键词, 核心画面, 构图建议, 产品占比, 场景元素, 图片文案, 禁止元素, 合规提醒, 英文生图 Prompt, 生成后检查清单, 实例演示.
- For the full brief template, see references/image-brief-template.md.
Image generation
- Generate images only after there is enough verified product structure and specs to avoid misleading visuals.
- Main image must have a pure white background and no text, person, props, badge, border, logo, or non-included accessories.
- Secondary images may use text, but all copy must be short, correctly spelled, and compliant.
- After generation, inspect images for product distortion, false specs, wrong text, logos, compliance risk, and whether a designer should refine them.
Deliverables
- Prefer a polished HTML report when the user asks for a shareable output.
- Include source links and a limitation note.
- When generating assets, save final files in the user’s requested output directory or current project outputs.

Bundled Assets

assets/sample-output/amazon_us_listing_audit_cosmo_report.html: a complete example report.
assets/sample-output/listing_images/: seven sample listing images generated from a filtered handheld shower head competitor brief.

Use these only as style and structure references. Do not copy their claims into another product unless the specs are verified for that product.

Amazon Listing COSMO

Core Rule

Never use competitor brand names in final title, bullets, backend terms, or image text. Use competitor names only in the audit tables when they are source data.

Workflow

Gather inputs

Product category, target marketplace, target product specs, brand restrictions, and competitor ASINs.
If ASINs or URLs are provided, browse/fetch current pages. Amazon content changes, so do not rely only on memory.
If the user asks for image generation, request or use verified product reference data. If unavailable, use only specs explicitly verified in the competitor/product source and disclose the limitation.

Competitor audit

Extract per ASIN: title, bullets, product details, A+ selling points, image sequence logic, review pain points, and high-frequency user scenarios.
Use 未获取 for missing fields.
Produce these tables: competitor basic info, selling point matrix, image expression matrix, review pain point matrix, borrowable points, avoid points.
For the detailed table pattern, see references/audit-template.md.

COSMO intent alignment

Build semantic shopping intent relationships, not just keyword stuffing:
- Product is_a
- Used_for_func
- Used_for_audience
- Used_in_location
- Used_on_body/object
- Capable_of
- Used_with
Use these relationships to connect product, problem, scene, audience, and accessory context.
For the scoring model, see references/cosmo-fabe-keywords.md.

FABE benefit rewriting

Convert parameters into user-perceived benefits:
- Feature: what it is
- Advantage: why it is better or easier
- Benefit: what the buyer feels or accomplishes
- Evidence: what image/detail/proof supports it
Avoid medical, absolute, unsupported, or safety claims.

High-quality listing output

Create keyword value score table.
Create Amazon title with highest-value terms in the first 80 characters.
Write bullets in this structure: CORE BENEFIT IN CAPS - scenario-based detail.
Provide English and Chinese bullets if the user asks in Chinese or requests bilingual output.
Include backend search terms and compliance exclusions.
For title/bullet patterns, see references/listing-copy-template.md.

7-image listing plan

Build exactly this sequence unless the user requests another:
1. Main Image
2. Lifestyle Image
3. Feature Infographic
4. Pain Point Solution
5. Compatibility / Size
6. Material / Detail
7. Use Case / Final Conversion
For each image, output: 图序, 图片类型, 转化目标, 对应用户痛点, 核心关键词, 核心画面, 构图建议, 产品占比, 场景元素, 图片文案, 禁止元素, 合规提醒, 英文生图 Prompt, 生成后检查清单, 实例演示.
For the full brief template, see references/image-brief-template.md.

Image generation

Generate images only after there is enough verified product structure and specs to avoid misleading visuals.
Main image must have a pure white background and no text, person, props, badge, border, logo, or non-included accessories.
Secondary images may use text, but all copy must be short, correctly spelled, and compliant.
After generation, inspect images for product distortion, false specs, wrong text, logos, compliance risk, and whether a designer should refine them.

Deliverables

Prefer a polished HTML report when the user asks for a shareable output.
Include source links and a limitation note.
When generating assets, save final files in the user’s requested output directory or current project outputs.

Bundled Assets

assets/sample-output/amazon_us_listing_audit_cosmo_report.html: a complete example report.

assets/sample-output/listing_images/: seven sample listing images generated from a filtered handheld shower head competitor brief.

Use these only as style and structure references. Do not copy their claims into another product unless the specs are verified for that product.