price-gap-monitor（价格差距监测）

Price Gap Monitor

Track visible price movement without pretending to know private marketplace data.

This skill now supports two operating modes under the same name.

Mode A — Product-level price trend monitoring

Use this mode when the user asks about:

• one specific product
• one brand-specific model
• one ASIN / listing / SKU
• one named product across multiple platforms

Mode B — Category price-band monitoring

Use this mode when the user asks about:

• a product category
• a keyword-defined market
• a visible price band
• cross-platform category pricing patterns

Browser-first guidance

When live public pages are available, prefer OpenClaw managed browser for page inspection.

Recommended order:

1. Use user-provided price snapshots if the user already has structured data.
2. If page URLs or searchable listings are available, use OpenClaw managed browser to inspect current public pricing and promo signals.
3. If the target marketplace gates pricing, ranking, or browsing depth behind login friction, explicitly remind the user to log in first so the agent can inspect fuller public results with fewer blockers.
4. Only use Browser Relay / attached Chrome when the user explicitly asks to inspect their current browser tab.

Do not default to Playwright-style assumptions in the user-facing guidance. The preferred browsing path is OpenClaw managed browser.

Login reminder rule

For marketplaces such as Amazon, trigger a login reminder when any of these conditions appear:

• search or category pages truncate, block, or degrade result visibility
• best-seller/category pages fail to load correctly
• location, cart, or account state is clearly affecting visible listings
• the task requires going deeper than a shallow guest snapshot

Suggested user-facing reminder:

• “If you want a cleaner and more complete Amazon read, log in first. Logged-in browsing usually gives more stable category pages, better listing continuity, and fewer interruptions.”

Do not claim login guarantees full data access. Present it as a practical way to improve visibility and continuity.

Core job

The goal is to produce a decision-ready price snapshot with honest trend interpretation.

This skill may use:

1. user-provided price snapshots, or
2. browser-collected public marketplace data

It should:

• collect visible price and promo signals
• compare listings or price bands
• distinguish current snapshot from repeated trend evidence
• recommend whether to watch, react, or gather more data first

It must not fabricate hidden marketplace history, real sales counts, or full coverage when only partial evidence is available.

Inputs

Input type A — user-provided snapshots

• competitor price tables
• prior exported marketplace snapshots
• your current price baseline
• target margin floor
• promo windows or campaign timing

Input type B — browser-collected public data

• a product model name
• an ASIN / SKU / listing URL
• a category keyword
• target platforms (Amazon, Temu, TikTok Shop, Walmart, etc.)
• market / locale (US, UK, JP, DE, etc.)

Workflow

Mode A — Product-level workflow

1. Define the exact product scope.
2. Collect visible public signals.
3. Normalize comparison points.
4. Determine evidence strength.
5. Produce result.

Mode B — Category-level workflow

1. Define the category scope.
2. Collect visible top listings.
3. Cluster the market.
4. Determine evidence strength.
5. Produce result.

Trend interpretation rules

1. Single snapshot rule

- If only one fresh snapshot is available, describe the result as a current price snapshot, not a full historical trend.

1. Repeated evidence rule

- Only describe an observed trend when supported by repeated visible price points or timestamped snapshots.

1. Sales honesty rule

- Never claim true sales volume unless the platform explicitly shows sold count. - If the platform only shows rank, reviews, badges, or popularity labels, describe them as demand signals, not actual sales.

1. Coverage rule

- If only part of the market is visible, clearly label the result as partial coverage. - Never present partial scraping as full category or full brand coverage.

1. History rule

- Never fabricate prior price history. - Never imply long-term movement when only current public pages were checked once.

Output format

For Mode A — product-level

1. Executive summary (max 5 lines)
2. Current product snapshot
3. Cross-platform comparison
4. Observed change or “insufficient trend history”
5. Risk / anomaly note
6. Recommended action (watch / act / gather more data)

For Mode B — category-level

1. Executive summary (max 5 lines)
2. Current category price-band snapshot
3. Platform comparison
4. Observed band shift or “insufficient trend history”
5. Noise vs real movement note
6. Recommended action (watch / act / gather more data)

Quality and safety rules

• Never recommend below the stated margin floor unless explicitly allowed.
• Avoid reacting to one-off noisy listing anomalies.
• Label uncertainty honestly.
• If browser results are thin or ambiguous, say so directly.
• Do not backfill missing marketplace data with guesses.

Creatop handoff

If the result is strong enough to act on, pass forward:

• accepted pricing actions
• watchlist items
• promo timing notes
• category price anchors
• cross-platform spread observations

License

Copyright (c) 2026 Razestar.

This skill is provided under CC BY-NC-SA 4.0 for non-commercial use. You may reuse and adapt it with attribution to Razestar, and share derivatives under the same license.

Commercial use requires a separate paid commercial license from Razestar. No trademark rights are granted.