Identify fraud patterns in your Stripe Radar dispute and refund data: look for correlated signals such as card country mismatch, velocity on a single IP, prepaid card usage, and new email domain patterns
In the Stripe Dashboard Radar Rules editor, create block rules using Radar's rule syntax; reference attributes such as card fingerprint velocity, IP risk score, email domain age, and shipping address mismatch
Create review rules for medium-risk signals that don't warrant an outright block but require manual inspection before fulfillment
Test rules in shadow mode (allow but flag) before promoting them to blocking mode; monitor false-positive rates on legitimate transactions
Pass custom metadata from your application to Stripe (e.g., account age, purchase history, device fingerprint) using the metadata field; reference it in Radar rules for richer signals
Monitor the Radar overview dashboard weekly and tune rule thresholds as fraud patterns evolve
Known gotchas
Overly aggressive Radar rules will block legitimate transactions; always measure and monitor false-positive rates alongside fraud block rates when tuning thresholds
Radar rules operate on the data available at request time; signals like account age or purchase history must be passed explicitly in metadata since Radar cannot query your database directly
Rule changes take effect immediately on new transactions; test in review mode before switching to block mode to avoid accidentally blocking high-value legitimate customers
Give your agent this knowledge — and 15,600+ more routes
One MCP install gives any agent live access to the full route map across 5,700+ domains, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp
Need this verified for your stack — or a route we don't have yet?