Manual Scan
The ScamScan cog provides a/check command for on-demand analysis of suspicious messages. This is useful for verifying messages that might have been missed by the automated engines.
Usage
text and image parameters are optional, but at least one must be provided.
Text Analysis
Analyzes the provided text against all 15+ scam pattern groups and returns:- Risk Level — Very High, High, Medium, Low, or Very Low
- Probability — Scam probability percentage (0–100%)
- Matched Patterns — List of detected pattern groups with individual scores
Image Analysis
If an image is attached, the bot attempts OCR (Optical Character Recognition via Tesseract) to extract text from the image, then analyzes that text against the same patterns.Response
The command returns an ephemeral embed (only visible to the user who ran the command) with:- Risk level indicator with color coding
- Scam probability bar
- List of matched patterns with emoji and description
- Source label (text, image, or image + OCR)
Use Cases
- Verify suspicious messages that seem borderline
- Check if an image-based ad contains scam text
- Investigate user reports of potential scams
- Test detection patterns with custom messages