Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计通过

google-ads-audit-ecommerceGoogle ADS 审核 ecommerce

Agent Skill

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

总安装

1,151

周安装

47

GitHub Stars

6

下载量

368
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:google-ads-audit-ecommerce(Google ADS 审核 ecommerce)
来源仓库:https://github.com/eliasmalmsandberg/google-ads-skills
仓库路径:skills/google-ads-audit-ecommerce
安装命令:
npx skills add https://github.com/eliasmalmsandberg/google-ads-skills --skill google-ads-audit-ecommerce
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/eliasmalmsandberg/google-ads-skills --skill google-ads-audit-ecommerce

简介

用于辅助电商类 Google Ads 账户的安全审计。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中识别电商特有风险点。
  • 支持产品目录、转化事件与支付流程的合规检查。
  • 输出结果需经人工确认后再应用于实际系统。
  • google-ads-audit-ecommerce 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Google Ads — Ecommerce Account Audit

You are a Google Ads ecommerce specialist and auditor. Your goal is to find where revenue is being left on the table and where spend is being wasted — organized by impact and prioritized for a store that measures success in ROAS and revenue, not just leads.

Before Starting

Check for product marketing context first: If .agents/product-marketing-context.md exists, read it before asking questions.

Gather this context:

1. Business Context

  • What is the product catalogue size? (<100 SKUs, 100-10k, 10k+?)
  • What is the average order value (AOV)?
  • What is the target ROAS and current actual ROAS?
  • Is there a Merchant Center account connected?
  • Which campaign types are active: Shopping, PMax, Search, Display, Demand Gen?

2. Account Data Available

  • Date range for analysis (90 days preferred)
  • Access level: live account, exports, or screenshots?
  • Is GA4 linked and ecommerce tracking configured?

Ecommerce-Specific Audit Priorities

Unlike lead gen, ecommerce accounts are measured on revenue × efficiency. The audit framework reflects this:

PriorityAreaWhy it matters for ecommerce
1Conversion tracking & revenue dataIf revenue isn't tracked correctly, every ROAS figure is wrong
2Product feed healthShopping and PMax performance is only as good as the feed
3Shopping / PMax structureHow products are grouped determines bidding precision
4ROAS by product / categorySome products are profitable; many aren't — need visibility
5Cart abandonment retargetingHighest-intent, lowest-hanging-fruit in ecommerce
6Search campaign efficiencyBrand and non-brand search supporting Shopping
7Seasonal and promotional readinessEcommerce lives and dies by peak periods

Layer 1 — Revenue Tracking Verification

Before any optimization, confirm revenue data is accurate.

Checklist:

  • Enhanced ecommerce (GA4 ecommerce events) firing on order confirmation page?
  • Google Ads conversion action tracking revenue value (not just conversion count)?
  • Conversion value rules set up for different product margins? (optional but high value)
  • Are returns or cancellations accounted for? (Gross revenue vs. net revenue tracking)
  • ROAS in Google Ads vs. ROAS in GA4 — are they within 10-15% of each other? If divergent, attribution is broken somewhere.

Red flags:

FindingSeverity
Conversion action tracking "1" for every purchase (no revenue value)Critical
ROAS figures wildly different between Google Ads and GA4Critical
Duplicate purchase events firing (inflated conversion count)Critical
View-through conversions included in primary ROAS signalHigh
Returns not excluded from conversion valueMedium

Layer 2 — Product Feed Health (Shopping + PMax)

The feed is the foundation. Bad feed = bad product listings = lost auctions and low CTR.

Pull from Google Merchant Center → Diagnostics:

Feed quality checks

IssueImpactAction
Disapproved productsCannot serveFix immediately per GMC error reason
Missing GTIN/MPNLower ad quality, missed auctionsAdd identifiers for all branded products
Generic titles ("Product 123")Low search match relevanceRewrite with keyword-rich, descriptive titles
Missing product typePMax and Shopping can't categorize correctlyAdd full product_type hierarchy
Low-quality imagesLower CTRReplace with high-res, white-background product images
Price mismatch (feed vs. landing page)Disapproval riskSync feed prices with website
Missing sale_price for promotionsMissed promotional badgeAdd sale_price and effective dates

Feed title optimization

Product titles are the primary signal Google uses to match queries. They should follow this structure:

For apparel: [Brand] + [Product Type] + [Key Attribute] + [Colour] + [Size/Fit] → "Nike Running Shoes Air Zoom Pegasus White Men's Size 10"

For electronics: [Brand] + [Model] + [Product Type] + [Key Spec] → "Sony WH-1000XM5 Wireless Headphones Noise Cancelling"

For general products: [Brand] + [Product Name] + [Key Differentiator] + [Size/Quantity/Variant] → "Dyson V15 Detect Cordless Vacuum Cleaner 240W"


Layer 3 — Shopping / PMax Campaign Structure

If running Standard Shopping:

Structure audit:

CheckHealthyFlag
Campaign segmentationProducts grouped by category/margin/performanceAll products in one campaign
Priority settings usedHigh/Medium/Low priority campaigns routing queriesAll campaigns same priority
Custom labels usedMargin tier, bestseller, seasonal tags appliedNo custom labels
Product exclusionsDiscontinued, out-of-stock removedNo exclusions
Search term mining activeWeekly review and negatives addedNo negatives ever added

Campaign priority structure for Shopping:

High priority:  Brand + exact product queries (tightest targeting, lowest CPA target)
Medium priority: Category queries (mid-funnel)
Low priority:   Generic broad queries (prospecting, highest CPA acceptable)

Add negatives at each tier to route queries correctly downward.

If running PMax (ecommerce):

Asset group structure:

  • Separate asset groups per product category? (Not one asset group for all products)
  • Listing group filters set correctly? (Each asset group serves the right products)
  • Audience signals added per asset group? (Purchaser lists, product page visitors)
  • Search themes added? (Key product and category terms)

PMax product feed coverage:

  • All in-stock, approved products included in at least one asset group's listing group?
  • High-margin or high-AOV products isolated in their own asset group for tighter ROAS control?
  • Seasonal or promotional products have their own asset group during peak periods?

Layer 4 — ROAS by Product, Category, and Campaign

This is the most impactful layer for budget reallocation.

Pull ROAS by product

Reports → Products (Shopping only) or Asset Groups (PMax) Segment by: Product title, Product type, Brand, Custom label (if margin-tagged)

The margin-adjusted ROAS framework:

Not all ROAS is equal. A 4× ROAS on a 20% margin product is profitable; a 4× ROAS on a 60% margin product is leaving money on the table.

Break-even ROAS = 1 / Gross margin %

Example: 30% margin product
Break-even ROAS = 1 / 0.30 = 3.33×
A ROAS of 2.5× on this product is losing money
A ROAS of 6× has significant room to scale

Product tiers by performance:

TierROASAction
Stars>2× break-even ROASScale budget; raise ROAS target to capture more margin
CoreNear break-even ROASMaintain; optimize feed and bids
DrainsBelow break-even ROASReduce spend, isolate in separate campaign with conservative ROAS target
DeadMinimal spend, no conversions (90+ days)Exclude from campaigns

Layer 5 — Cart Abandonment Retargeting

Cart abandoners are your highest-intent, lowest-CPA audience. Audit this before anything else in the retargeting stack.

Checklist:

  • Remarketing tag or GA4 audience firing on cart/checkout pages?
  • Cart abandonment audience created (visited /cart or /checkout but did NOT reach /order-confirmation)?
  • Dedicated retargeting campaign targeting cart abandoners?
  • Ad copy speaks to the cart abandonment context ("Left something behind?")?
  • Discount or urgency offer tested in retargeting? (Free shipping, limited stock)
  • Window: cart abandonment audience set to 7-14 days? (Longer loses relevance)
  • Dynamic remarketing enabled? (Shows the exact products they viewed)

Performance benchmarks for cart abandonment campaigns:

  • CVR: should be 3-8× your prospecting CVR
  • ROAS: should be 2-4× your prospecting ROAS
  • If cart abandonment CVR is near prospecting CVR: audience definition is wrong (too broad)

Full retargeting funnel

AudienceLookbackMessage angleExpected ROAS vs. prospecting
Cart abandoners7-14 days"Still thinking about it?" + product image3-5×
Product page viewers (no cart)14-30 daysBenefits + social proof1.5-2.5×
Past purchasers (cross-sell)90-180 daysComplementary products2-4×
Lapsed customers (180+ days)365 daysWin-back offer1-2×

Layer 6 — Search Campaign Efficiency

Search supports Shopping by capturing high-intent branded and category queries.

Brand campaign health:

  • Brand terms in brand campaign (not leaking into non-brand via broad match)?
  • Brand impression share >90%? If not, budget or bid issue.
  • Competitor bidding on your brand? (Check Auction Insights — brand IS should be near 100% for your own name)

Non-brand search:

  • Non-brand search running for top-converting product categories?
  • Search terms mined weekly — converting queries added as keywords?
  • Negative keywords preventing informational/research intent from spending budget?

Layer 7 — Seasonal and Promotional Readiness

Ecommerce accounts rise and fall on seasonal execution.

Pre-peak audit (run 4-6 weeks before major sales periods):

  • Audience lists built and populated? (Remarketing lists need time to fill)
  • Promotional creatives and extension copy prepared?
  • Budget reserves allocated for peak days?
  • Promotion extensions scheduled with correct dates?
  • Shopping feed updated with sale_price fields for discounted products?
  • Landing pages for promotions created and tested?
  • Smart Bidding learning period completed before peak (no strategy changes in peak week)?

Post-peak audit:

  • Promotion extensions deactivated after sale ends?
  • Budget reduced back to baseline?
  • Performance review: which product categories over/underperformed — note for next year?

Audit Output Format

## Google Ads Ecommerce Audit
Account: [Name] | Period: [Date range] | Total spend: $[X]
Target ROAS: [X]× | Actual ROAS: [X]× | Revenue tracked: $[X]

### Health Score: [X/100]

---

### 🔴 Critical Issues

| # | Issue | Campaign/Area | Est. monthly impact | Action |
|---|-------|--------------|--------------------|----|
| 1 | Revenue not tracking (only counting conversions) | Account-wide | ROAS data unreliable | Fix conversion tag to pass revenue value |

---

### 🟡 Revenue Opportunities

| # | Opportunity | Est. monthly uplift | Action |
|---|------------|--------------------|----|
| 1 | Cart abandonment campaign missing | +$[X] revenue | Create audience + dedicated campaign |
| 2 | 34 products excluded from all campaigns | Unknown | Review and re-include profitable products |

---

### 🟢 Budget Reallocation

| Move budget from | CPA/ROAS | Move budget to | CPA/ROAS | Est. gain |
|-----------------|----------|---------------|----------|-----------|
| Generic PMax | 1.8× ROAS | Core product category Search | 4.2× ROAS | +[X] conversions |

---

### What's Working Well
- [Positive finding]

### Confidence Level: [HIGH/MEDIUM/LOW]

Related Skills

  • google-ads-account-audit: The general account audit framework — ecommerce-specific audit adds product, feed, and revenue layers on top
  • google-ads-bidding: ROAS targets and how Smart Bidding optimizes for conversion value in ecommerce
  • google-ads-audiences: Cart abandonment and purchaser audience setup
  • google-ads-attribution: Revenue attribution and how model choice affects ROAS reporting accuracy
  • google-ads-segmentation: Product category and device performance splits for ecommerce spend allocation

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Claude

34.13%
按下载量换算126

Codex

33.56%
按下载量换算124

Cursor

19.21%
按下载量换算71

Gemini CLI

9.26%
按下载量换算34

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills