In the Stripe Dashboard navigate to Radar > Rules and understand the three rule actions: block, review, and allow.
Create a block rule using the risk_score attribute (e.g., ':risk_score: > 85') to stop high-confidence fraud before authorization.
Chain a review rule at a lower threshold (e.g., ':risk_score: > 65') so borderline charges land in the review queue rather than being silently passed.
Add metadata-aware conditions using custom attributes — attach metadata fields on the PaymentIntent and reference them in rules with the 'metadata' namespace.
Use the 'is_3d_secure' attribute to skip stricter rules for SCA-authenticated payments, reducing false positives on challenged transactions.
Test rules in 'shadow mode' (simulate without enforcing) before promoting to live, then monitor the Radar review queue for false-positive rates.
Known gotchas
Rule evaluation order matters: Stripe evaluates rules top-to-bottom and stops at the first match, so place narrow allow rules above broad block rules or they will never fire.
Custom metadata attributes must be present on the PaymentIntent at creation time; metadata added after charge creation is not visible to Radar rule evaluation.
Overly aggressive score thresholds on cross-border or high-value segments can spike false-positive rates and trigger issuer friction — monitor authorization rates by segment after any rule change.
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp