with one click
product-selling-point-analysis
// 分析产品图片、产品名和描述,提炼卖点、目标人群、痛点、使用场景和内容角度。适合“小程序产品库的产品分析”“帮我分析这个产品卖点”“从图片提炼卖点”等请求。
// 分析产品图片、产品名和描述,提炼卖点、目标人群、痛点、使用场景和内容角度。适合“小程序产品库的产品分析”“帮我分析这个产品卖点”“从图片提炼卖点”等请求。
[HINT] Download the complete skill directory including SKILL.md and all related files
| name | product-selling-point-analysis |
| description | 分析产品图片、产品名和描述,提炼卖点、目标人群、痛点、使用场景和内容角度。适合“小程序产品库的产品分析”“帮我分析这个产品卖点”“从图片提炼卖点”等请求。 |
| allowed-tools | Read, Write, Bash |
Follow shared NexTide rules in:
nextide-sharedUse this skill when the user wants to:
Do not use this skill for:
sourcing-selectionxiaohongshu-note-collectorCapability id:
product.selling_point.analysis
CLI contract:
npm run nextide -- capability run product.selling_point.analysis \
--input .nextide/input/product-selling-point-analysis.json \
--output .nextide/output/product-selling-point-analysis-result.json \
--mode wait
For local debugging only, credentials may be supplied via config/env/flag. Prefer nextide auth login or NEXTIDE_USER_API_KEY=<redacted> over putting keys in commands.
Auth note:
--user-api-key <key> or NEXTIDE_USER_API_KEY=<key>.apiKey / api_key in input is only a backward-compatible fallback.unauthorized.{
"name": "家用射频美容仪",
"description": "产品描述...",
"images": ["https://..."]
}
Field rules:
name is required.description is recommended.images should include at least one product image URL when visual analysis matters.apiKey in the JSON fixture; pass it through CLI/env/config.Collect product name, description, and image URLs from the user.
Ask for missing essentials only when the analysis would be misleading; otherwise run a bounded first pass.
Create .nextide/input/product-selling-point-analysis.json with business input only, no credentials.
Run the capability command.
If a runId exists, export the artifact/data bundle:
RUN_ID=$(node -e "const r=require('./.nextide/output/product-selling-point-analysis-result.json'); console.log(r.run && r.run.runId)")
npm run nextide -- run artifacts "$RUN_ID" \
--output-dir .nextide/output/$RUN_ID \
--download \
--gallery \
--datatable
Read summary.json and datatable.json first when present.
Summarize into user-facing product insight, separating facts from hypotheses.
Expected successful result shape:
{
"run": {
"status": "succeeded",
"result": {
"sellingPoints": [],
"detailedDescription": "...",
"workflowData": {}
}
}
}
If status is waiting_callback, report that the n8n flow accepted the job but full structured results may arrive asynchronously.
Return:
datatable preview for the product insight table when generatedOutput format preference:
datatable.json exists.explanation, stop and give the user the recommended next actions.product.selling_point.analysis0.2.0productavailableinternal_apifalselownexTideApiKeynonenonenoneproduct, selling-points, analysisDescription:
分析产品图片与描述,提炼卖点、目标人群、痛点和内容角度。
Examples:
产品卖点分析
{
"name": "护颈枕",
"description": "适合久坐人群的支撑枕",
"images": []
}
Input fields:
name (string, required):产品名称。description (string):产品描述。images (string[]):产品图片 URL 列表。productId (string):已有产品 ID。可选。Output fields:
sellingPoints (string[]):卖点列表。detailedDescription (string):详细产品分析文本。workflowData (object):完整结构化结果。CLI:
nextide capability run product.selling_point.analysis \
--input .nextide/input/product.selling_point.analysis.json \
--output .nextide/output/product.selling_point.analysis-result.json \
--mode wait
If the result contains artifacts, export them:
RUN_ID=$(node -e "const r=require('./.nextide/output/product.selling_point.analysis-result.json'); console.log(r.run && r.run.runId)")
nextide run artifacts "$RUN_ID" \
--output-dir .nextide/output/$RUN_ID \
--download \
--gallery \
--datatable
Then read summary.json first, followed by manifest.json.
available, fail fast and explain what is missing.--wait when the user wants a finished result in the same turn.nextide run artifacts <run-id> --output-dir .nextide/output/<run-id> --download --gallery --datatable and read summary.json then manifest.json.nextide run follow <run-id> --output-dir .nextide/output/<run-id> --timeout 1800 --interval 5.summary.recommendedResponse.message, preview.html, datatable.json, and local artifact paths over pasting huge raw JSON.explanation, convert it into a clear user-facing failure message with next actions.