| name | ecommerce.trend-discovery |
| version | 2.0.0 |
| description | Discover rising ecommerce categories and emerging niches across platforms. Triggers: what is trending, trending products, rising categories, next big thing. Use for trend discovery, not validating a specific niche or finding new Amazon launches.
|
| allowed-tools | ["Bash","Read","Write"] |
IMPORTANT — Working Directory: All scripts in this skill MUST be executed with the working directory set to the absolute path of the directory containing this SKILL.md file (the skill root directory). Always cd into this directory before running any command. Do NOT use relative paths or shell shortcuts like ~ in paths, as they may not be resolved by the execution environment.
Trend Discovery v2.0.0
Category / Market Understanding Step
Before running this skill, if the user provides a broad product category, niche, keyword, market idea, or trend topic, first identify:
- category or niche
- target marketplace, country, or platform
- user's analysis goal
- relevant seed keywords or subcategories
- whether the request is about demand, competition, trend, opportunity, or prioritization
Use this market understanding to choose keywords, filters, regions, comparison scope, and analysis dimensions before executing the script.
What categories are trending? What niches are rising?
Discover rising e-commerce categories using cross-platform data analysis.
Core Question
— What categories are trending?
— What's hot on TikTok?
When to Use
- Starting product research and need direction
- Finding rising niches before they become competitive
- Validating category momentum across platforms
- Comparing trend strength between keywords
- Identifying seasonal vs sustained trends
Differs From / Not Applicable
- Use market-overview after selecting a market/category to understand it deeply.
- Use new-product-tracker for newly launched Amazon products.
- Use niche-evaluator for go/no-go entry scoring.
- Use this skill for early trend/category discovery.
Data Sources
| Source | Endpoint | Data |
|---|
| TikTok Echotik | /api/v1/tools/research/echotik/listProduct | Hot products, sales velocity |
| Google Trends | /api/v1/tools/research/googleTrend/getTrendByKeys | Search interest over time |
| Amazon Search | /api/v1/tools/research/amazon/search | Product listings, competition |
| Jungle Scout | /api/v1/tools/jungle-scout/keywords/historical-search-volume (POST) | 30/90/180-day search volume trends |
Analysis Modes
| Mode | Use Case |
|---|
+cross-platform | Full discovery (TikTok → Google → Amazon) |
+tiktok-hot | TikTok trending products only |
+compare | Compare specific keywords head-to-head |
+timeline | Single keyword historical trend |
⚠️ MANDATORY: Charts & Display Rules
- Generate charts by default for full analytical reports — use
--chart with an output directory unless the user asks for a quick text-only result or chart data is unavailable
- Send generated charts to the user — when charts are created, share all chart PNGs immediately
- All chart styling is driven by
scripts/chart_style.json (derived from references/display-rules.md). If you need to change colors/fonts/styles, edit chart_style.json — do NOT hardcode styles in the script.
- To update chart_style.json: read
references/display-rules.md, extract the relevant values, write to scripts/chart_style.json.
Usage
Always include --chart /tmp/<descriptive-name> in every run:
python3 scripts/trend_discovery.py '{"keywords": ["yoga mat", "pilates mat"]}' --chart /tmp/trend-charts
python3 scripts/trend_discovery.py '{"keywords": ["kombucha"]}' --chart /tmp/trend-charts
After the script finishes, send all generated chart PNGs to the user.
Parameters
| Parameter | Type | Default | Description |
|---|
mode | string | "cross-platform" | Analysis mode |
region | string | | Target region |
keywords | array | - | Keywords to compare |
keyword | string | - | Single keyword for timeline |
min_sales | int | 5000 | Minimum 30d sales for TikTok |
days | int | 90 | Historical days for trends |
Output Structure
The output will be a structured markdown report, following this format:
Trend Discovery Report: [Keywords]
1. Executive Summary
- Analyzed Keywords: [analyzed_keywords]
- Analysis Mode: [mode] (for example: cross-platform, TikTok hot, comparison, timeline)
- Confidence Level: [insights.confidence_level]
- Key Insight: [insights.summary]
2. Trend Overview
| Trend Category | Keyword Count |
|---|
| 🔥 Hot | [insights.trend_distribution.hot] |
| 📈 Rising | [insights.trend_distribution.rising] |
| ➡️ Stable | [insights.trend_distribution.stable] |
| 📉 Declining | [insights.trend_distribution.declining] |
| Platform | Count (Keywords with Signals) |
|---|
| TikTok Hot | [insights.platform_signals.tiktok_hot] |
| Google Rising Trend | [insights.platform_signals.google_rising] |
| Amazon New Product Success | [insights.platform_signals.amazon_new_success] |
3. Discovered Trends
- Keywords sorted by trend score (show top 5):
| Rank | Keyword | Trend Score | Trend Verdict | Signal Sources | Google Trend Direction | Google Trend Change % | TikTok Sales (Estimated) | Amazon New Product Success |
|---|
| 1 | [trends[0].keyword] | [trends[0].trend_score.score] | [trends[0].trend_score.verdict] | [trends[0].trend_score.signals] | [trends[0].google_trends.direction] | [trends[0].google_trends.change_pct]% | [trends[0].tiktok.total_sales] | [trends[0].amazon.new_entrant_success] |
| ... | ... | ... | ... | ... | ... | ... | ... | ... |
4. Overall Recommendations
- All Recommendations:
- [insights.recommendations[0]]
- [insights.recommendations[1]]
- ...
5. Attached Visualizations
- Trend Score Comparison (trend_comparison.png)
- Google Trends (google_trends.png)
- Market Maturity (market_maturity.png)
- Score Breakdown (score_breakdown.png)
Trend Classification
| Signal | Criteria | Action |
|---|
| 🔥 HOT | TikTok viral + Google rising + Amazon growing | Enter quickly |
| 📈 RISING | Consistent upward trend across 2+ platforms | Good opportunity |
| ➡️ STABLE | Flat trend, established market | Evaluate competition |
| 📉 DECLINING | Downward trend | Avoid or exit |
| 🎭 HYPE | TikTok hot but no Amazon demand | Risky, may not convert |
Charts Generated
| Chart | Description | File |
|---|
| Trend Score Comparison | Horizontal bar chart of keyword scores | trend_comparison.png |
| Google Trends | Multi-keyword search interest over time | google_trends.png |
| Market Maturity | Pie chart of emerging/growing/mature split | market_maturity.png |
| Score Breakdown | Top keyword score composition | score_breakdown.png |
Cross-Platform Validation
| Platform | Signal Weight | Why |
|---|
| TikTok | 35% | Early trend indicator, viral potential |
| Google Trends | 30% | Consumer interest validation |
| Amazon | 35% | Actual purchase intent |
Workflow Integration
Discovery Phase
|-- trend-discovery -> Find trending categories -> YOU ARE HERE
|-- market-overview -> Understand the market
`-- niche-evaluator -> Evaluate niche potential
Environment Variables
| Variable | Required | Description |
|---|
NEXSCOPE_API_KEY | Yes | NexScope proxy API key |
NEXSCOPE_PROXY_BASE | Yes | NexScope proxy base URL |
Limitations
- TikTok data limited to supported regions
- Google Trends has 1-2 day lag
- Cross-platform correlation is approximate
- Some trends are regional only
- Viral products may not have long-term demand
References
references/api-reference.md - API endpoint documentationreferences/display-rules.md - Chart styling guidelines (source of truth)scripts/chart_style.json - Machine-readable chart style config (derived from display-rules.md, loaded by script at runtime)
Multi-Batch Usage
When analyzing more items than you want to run in a single invocation, keep charts and reports aligned by saving each batch as raw JSON and then generating one merged result.
Step 1: run each batch with --output to save intermediate JSON.
python3 scripts/trend_discovery.py '{"keyword": "example"}' --output /tmp/batch1.json
python3 scripts/trend_discovery.py '{"keyword": "example 2"}' --output /tmp/batch2.json
Step 2: merge every batch JSON and generate the final unified chart.
python3 scripts/trend_discovery.py --merge /tmp/batch1.json /tmp/batch2.json --sort score --chart /tmp/final-charts
Use the merged JSON output and /tmp/final-charts/merged_ranking.png for the final report. Do not present per-batch charts as final charts when the text report has been globally re-ranked.
