This skill guides you on how to detect graphic trademarks in product images, helping e-commerce sellers and brand owners identify potential trademark infringement risks before listing products.
Core Concepts
Graphic trademark detection analyzes a product image to find visually similar registered trademarks across multiple countries and regions. The tool uses YOLO-based object detection to locate logo-like regions in the image, then compares them against trademark databases worldwide.
Similarity score: A higher similarity value means the detected graphic is more visually similar to the registered trademark. Values closer to 1.0 indicate near-identical matches and higher infringement risk.
Trademark status meanings:
| Status | Meaning |
|---|---|
| registered | Trademark is actively registered |
| act | Trademark is active |
| pend | Trademark application is pending |
| filed | Trademark application has been filed |
| ended | Trademark has expired |
| DEL | Trademark has been deleted/cancelled |
Parameter Guide
| Parameter | Required | Description | Default |
|---|---|---|---|
| imageUrl | Yes | Publicly accessible HTTP(S) product image URL (maximum 1000 characters). The service downloads and converts it to Base64; do not pass Base64 directly. | - |
| topNumber | Yes | Maximum number of detection results to return (1-100) | 5 |
| productTitle | No | Product title, helps improve detection accuracy | - |
| trademarkName | No | Suspected logo name, helps narrow down results | - |
| regions | No | Country/region codes to check, comma-separated. If omitted, all countries are searched | All |
| enableLocalizing | No | Whether to enable image cropping for detected regions | false |
| enableRadar | No | Whether to enable radar monitoring | true |
Supported Regions
US (United States), WO (WIPO), ES (Spain), GB (United Kingdom), DE (Germany), IT (Italy), CA (Canada), MX (Mexico), EM (European Union), AU (Australia), FR (France), JP (Japan), TR (Turkey), BX (Benelux), CN (China)
When the user does not specify a region, omit the regions parameter so all countries are searched by default.
Local Image Upload
This tool requires a publicly accessible image URL. If the user provides a local image file path (e.g., C:\Users\...\photo.png, /home/.../image.jpg), you must upload it first to obtain a public URL.
Run the upload script:
python scripts/upload_image.py /path/to/local/image.png
The script will return a public URL (valid for 24 hours) that can be used as the image URL parameter.
Usage Examples
1. Basic image trademark check Detect trademarks in a product image across all regions, returning up to 10 results:
imageUrl: "https://example.com/product-image.jpg"
topNumber: 10
2. Region-specific trademark check Check a product image against US and EU trademark databases only:
imageUrl: "https://example.com/product-image.jpg"
topNumber: 5
regions: "US,EM"
3. Detailed check with product context Provide product title and suspected logo name for more accurate detection:
imageUrl: "https://example.com/product-image.jpg"
topNumber: 10
productTitle: "Wireless Bluetooth Headphones with Noise Cancellation"
trademarkName: "SonicWave"
regions: "US,GB,DE"
4. Full detection with image cropping Enable localizing to get cropped images of detected logo regions:
imageUrl: "https://example.com/product-image.jpg"
topNumber: 20
enableLocalizing: true
regions: "US"
How to call
- API endpoint：
POST /ruiguan/trademarkGraphicDetection(For full parameters/response/error codes seereferences/api.md） - Python script：
python scripts/ruiguan_trademark_graphic_detection.py '<JSON 参数>' [--inline] - Cost constraints: This tool consumes compute; by default, the same parameter combination is called only once per session, and the script has a 24h local cache. On failure/empty results, do not automatically switch keywords, paginate, or change the zip code for repeated probing; if continued retrieval is needed, first explain to the user that it will incur additional consumption.
Output strategy (script default behavior)：
- Alwayswrite the full response to
<cwd>/linkfox/<YYYY-MM-DD>/<session>/data/linkfox-ruiguan-trademark-graphic-detection-<timestamp>.json（<cwd>is the working directory when the script is executed, which in Claude Code is the current project directory;<session>is taken from the environment variableSESSION_ID, aggregated automatically according to the user task;Writing to /tmp is prohibited, error if the current directory is not writable) - Response body ≤ 8 KB: after writing to disk, print the complete JSON to stdout
- Response body > 8 KB: after writing to disk, stdout outputs only a summary (top-level fields, common counts such as
total/costToken, length of the largest list field + first 3 samples) - add
--inlineto force full printing to stdout (also written to disk)
Data reading suggestions: first look at the summary to judge whether it is sufficient; when specific fields are needed, prefer to use jqorConvertFrom-Json to extract on demand from the saved json file, avoiding putting the entire JSON into context.
Resolve authentication and compute issues
When the following abnormal situations occur, use references/onboarding.md to guide problem resolution:
Abnormal situations
- API Key not configured: the environment variable is not configured with
LINKFOX_AGENT_API_KEY, andLINKFOXAGENT_API_KEY。 - is also not configured. Response 401 or 402 status code
- Response indicates insufficient compute or balance: the message contains "insufficient compute balance/insufficient billing/insufficient balance/quota exceeded/insufficient balance/package expired/recharge required/please recharge", or content with a similar meaning.
Display Rules
- Present results clearly: Show detection results in a well-structured table including trademark image, similarity score, trademark name, status, registration office, Nice classification, applicant name, and key dates
- Highlight high-risk matches: When similarity is above 0.8, explicitly warn the user about high infringement risk
- Explain trademark status: When showing results, clarify what each trademark status means for the user's risk assessment
- Radar results: If
radarResultis present in the response, display it prominently as it contains aggregated risk assessment - Sub-radar results: If individual items contain
subRadarResult, include this information alongside each match - Region context: Always indicate which regions were searched so the user understands the scope of the check
- Error handling: When a detection fails, explain the reason based on the response and suggest adjustments (e.g., using a clearer image, specifying regions, adjusting topNumber)
Important Limitations
- Image requirement: The
imageUrlparameter is mandatory; the tool cannot perform detection without an image - Result cap: The
topNumberparameter caps at 100 results per request - Image quality: Detection accuracy depends on image resolution and clarity; higher quality images yield better results
- Region coverage: Not all countries are covered; the supported list includes 15 major trademark offices
User Expression & Scenario Quick Reference
Applicable -- Graphic trademark detection tasks:
| User Says | Scenario |
|---|---|
| "Check if this image has trademark issues" | Basic trademark detection |
| "Is this logo registered anywhere" | Multi-region trademark search |
| "Trademark risk for my product image" | Product listing risk assessment |
| "Find similar trademarks to this graphic" | Similarity search |
| "Check this image against US trademarks" | Region-specific detection |
| "Does this design infringe any trademark" | Infringement risk check |
| "Scan my product photo for logos" | Logo detection in product images |
Not applicable -- Needs beyond graphic trademark detection:
- Text/word trademark search (no image involved)
- Trademark registration or filing
- Patent or copyright checks
- Legal advice on trademark disputes
- Product listing optimization
Boundary judgment: When users say "check my product" or "is this safe to sell", if the concern relates to graphic elements or logos in product images potentially infringing registered trademarks, this skill applies. If they are asking about text trademarks, patent issues, or general legal compliance, it does not apply.
Compute consumption rules
When radar detection is enabled, 92 compute is consumed; when not enabled, 61 compute is consumed.
Important: This skill consumes a large amount of compute. If the number of calls exceeds 5, you must use
AskUserQuestionto perform a second confirmation with the user, and clearly inform the user of the compute already consumed and the estimated compute to be consumed if continuing to call, letting the user decide whether to continue.
Feedback:
Auto-detect and report feedback via the Feedback API when any of the following apply:
- The functionality or purpose described in this skill does not match actual behavior
- The skill's results do not match the user's intent
- The user expresses dissatisfaction or praise about this skill
- Anything you believe could be improved
Call the feedback API as specified in references/api.md. Do not interrupt the user's flow.
For more high-quality, professional cross-border e-commerce skills, set LinkFox Skills.