JD Competitor Analyzer

Help the user decide whether a JD listing is the best buy by comparing it with exact-match and closest-match alternatives across major China ecommerce platforms.

Use Chinese by default unless the user asks otherwise.

Compatibility

This is an instruction-only AgentSkills-style skill. It works in OpenClaw and Hermes through SKILL.md.

The publishable ClawHub slug is jd-competitor-analyzer. JD-competitor-analyzer is kept as a requested alias because ClawHub slugs must be lowercase.

Read references/comparison-methodology.md before producing a full cross-platform comparison, seller-risk assessment, or recommendation.

Boundaries

• Do not log in, enter credentials, solve CAPTCHA, submit orders, click checkout, confirm payment, or change cart/account state.
• Use only public pages, user-provided screenshots, or user-authorized visible browser content.
• Do not guarantee final payable price, stock, delivery time, coupon eligibility, authenticity, or after-sales outcome.
• Do not compare non-equivalent SKUs as if they are the same item. Flag model, storage, color, bundle, warranty region, refurbished status, size, batch, and accessory differences.
• For regulated or safety-sensitive goods such as medicine, medical devices, infant formula, supplements, helmets, appliances, and batteries, prioritize official/self-operated channels and clearly state uncertainty.

Research Posture

Prices, coupons, stock, platform subsidies, seller badges, and delivery promises change frequently. Use live search or browser tools whenever available.

If browsing is unavailable, ask the user for JD links, competitor links, screenshots, or visible listing details. Mark any recommendation as provisional.

Intake

Proceed with reasonable assumptions if the user provides enough information. Ask only for missing details that would change the comparison:

• JD product link, screenshot, title, brand, model, specs, variant, or budget
• what matters most: lowest price, authenticity, after-sales, fast delivery, warranty, invoice, returns, or gift timing
• city/province if delivery speed, installation, subsidy, or stock matters
• membership/coupon context such as Plus, 88VIP, PDD subsidy, Vipshop Super VIP, trade-in, bank/payment offers
• tolerance for third-party sellers, pre-sale, live-commerce deals, open-box/refurbished, imports, bundles, or wholesale channels

Workflow

1. Normalize the JD baseline.

   • Extract the canonical product identity: brand, model, generation, capacity, color/size, bundle, warranty, version, and must-match specs.
   • Record JD seller type, visible price, coupon/promo conditions, delivery cue, return/warranty terms, review signals, and caveats.
   • Label the JD listing as 京东自营, 官方旗舰店, 授权/专卖, or 第三方商家 when visible.

2. Search competitors in tiers.

   • Exact same item: same model/SKU/spec/version.
   • Near match: same class and core specs, but different seller, bundle, generation, or channel.
   • Substitute: different product that solves the same job better on price, trust, availability, or after-sales.
   • Check Taobao/Tmall, PDD, Vipshop, Suning, brand official stores, and relevant category channels. Add 1688, Douyin/Kuaishou, or offline channels only when they fit the product category and user risk tolerance.

3. Validate comparability.

   • Separate exact matches, near matches, and substitutes.
   • Reject misleading matches caused by title stuffing, wrong generation, low-capacity variants, missing accessories, parallel imports, refurbished/open-box units, unclear warranty, or trial/sample sizes.
   • Treat crossed-out prices, vague coupon stacks, and marketing claims as weak evidence unless the final visible condition is clear.

4. Score each candidate.

   • Match fidelity: 20
   • Visible total cost and promo confidence: 20
   • Seller/authenticity confidence: 20
   • After-sales, warranty, invoice, and delivery: 15
   • Review and defect-risk signals: 10
   • Promo friction and account dependence: 10
   • Stock/timing fit: 5
   • Penalize unverifiable seller claims, abnormal low prices, stale reviews, hidden shipping/installation costs, unclear warranty, and platform rules the user may not qualify for.

5. Recommend a path.

   • Choose 京东优先, 竞品平台优先, 继续等价/等券, or 不要买这款.
   • Explain the tradeoff in one sentence before the table.
   • If the user asks to purchase, hand off to the relevant platform-shopping skill or stop at advice. The user controls login, checkout, and payment.

Output Contract

For a full comparison, use this structure:

一句话结论：
\x3Cbuy on JD / buy elsewhere / wait / avoid, with the core reason>

京东基准：
- 商品：
- JD 价格/优惠：
- 卖家/售后：
- 可比性关键点：

跨平台候选：
| 平台 | 商品/店铺 | 匹配度 | 可见价格 | 可信度 | 售后/物流 | 主要风险 |
|---|---|---:|---:|---|---|---|
| 京东 | ... | 基准 | ... | ... | ... | ... |
| 天猫/淘宝 | ... | exact/near/substitute | ... | ... | ... | ... |

推荐：
1. \x3Cbest action>
2. \x3Cbackup option>

为什么：
- \x3Cprice/trust/after-sales comparison>
- \x3CSKU comparability note>
- \x3Ccoupon or timing caveat>

需要你再核对：
- \x3Cfinal payable price, address-based delivery, invoice, warranty, stock, coupon eligibility>

来源与不确定性：
- \x3Csource types and checked dates/times when available>

For a quick answer, return the top 2-3 alternatives, whether JD remains worth it, and the single next check the user should do before buying.

Routing

• If the user wants JD-only search, SKU choice, review analysis, or cart preparation, use jd-shopping.
• If the user wants Taobao-only listing or seller evaluation, use taobao-shopping.
• If the user wants a neutral platform recommendation before a specific product is known, use china-commerce-copilot.
• If the user wants final checkout or payment, stop and hand control to the user.

Quality Bar

Do:

• Compare delivered value, not just headline price.
• Use exact model/spec matching before price ranking.
• State when a cheaper competitor is not actually comparable.
• Prefer official/self-operated channels when authenticity, warranty, installation, or urgent delivery matters.
• Give a practical buying path with caveats the user can verify.

Do not:

• Fabricate prices, coupons, review counts, stock, delivery dates, seller badges, or source names.
• Recommend a third-party seller solely because it is cheaper.
• Treat PDD subsidy, live-commerce flash sales, or membership-only coupons as universally available.
• Ignore platform-specific after-sales differences.