# Screen Transactions

Fordefi can screen outgoing transactions with external screening providers —
Blockaid and Hypernative — before they are signed, and return the results
through the API alongside Fordefi's own risk engine output. Screening is opt-in
per organization: an admin activates a provider by saving valid credentials for
it in **Settings** > **Screening** in the web console. See
[Integrate transaction screening providers](/user-guide/transaction-screening)
for the setup and the supported chains.

Once a provider is active, its findings reach you in two forms:

- **Risk objects**, in the existing `risks` array, so screening findings arrive
through the same field your integration already reads.
- **Screening objects**, in a `screening` array, which report the per-provider
outcome — including a clean result — and carry the provider's own response
details.


## Risk objects

Screening findings are turned into risk objects, one per finding:

- **Blockaid:** One risk per entry in `validation.features` whose `type` is not
`Benign`.
- **Hypernative:** One risk per entry in `findings` whose `severity` is neither
`accept` nor `autoAccept`.


## Screening objects

Every endpoint that returns `risks` also returns a `screening` array, with one
entry per external provider that the transaction was screened with. Fordefi's
own risk engine is not represented in `screening`. The array may be empty — for
example when no provider is active for the organization.

| Field | Type | Description |
|  --- | --- | --- |
| `vendor` | string enum | `blockaid` or `hypernative`. |
| `result` | string | Blockaid: `Malicious`, `Warning`, or `Benign`, from `validation.result_type`. Hypernative: `accept`, `autoAccept`, `warn`, or `deny`, from `data.recommendation`. |
| `screening_details` | object | The relevant part of the provider's raw response. See [Screening details](#screening-details). |


Read `screening` rather than inferring the outcome from `risks`: a provider that
returned a clean result and a provider that was never called both produce zero
risks, and only `screening` tells them apart.

### Screening details

`screening_details` carries through the provider's own response, so you can
correlate a Fordefi transaction with a screening on the provider's platform.

### Examples

A Blockaid screening object:

```json
"screening": [
  {
    "vendor": "blockaid",
    "result": "Malicious",
    "title": "known_malicious",
    "description": "A known malicious address is involved in the transaction",
    "screening_details": {
      "validation": {
        "status": "Success",
        "result_type": "Malicious",
        "description": "A known malicious address is involved in the transaction",
        "reason": "b",
        "classification": "known_malicious",
        "features": [
          {
            "type": "Malicious",
            "feature_id": "KNOWN_MALICIOUS_ADDRESS",
            "description": "This address has recorded malicious activity",
            "address": "0x0a5738064da9dda3cf4a4ae8d8e49b16fd7467c5"
          }
        ]
      }
    }
  }
]
```

A Hypernative screening object:

```json
"screening": [
  {
    "vendor": "hypernative",
    "result": "deny",
    "title": "Phishing victim - approval",
    "description": "[Simulated] An address approved funds to a phishing/scamming address",
    "screening_details": {
      "data": {
        "recommendation": "deny",
        "assessmentId": "fdbf8e7b-8387-4fbf-b3fc-3eeb9ca39d54",
        "assessmentTimestamp": "2026-01-21T08:01:18.375Z",
        "findings": [
          {
            "typeId": "S-1925",
            "title": "Phishing victim - approval",
            "description": "[Simulated] An address approved funds to a phishing/scamming address",
            "severity": "Deny"
          }
        ]
      },
      "error": null
    }
  }
]
```

## Timeouts and error handling

How the policy engine treats an unusable screening result depends on why it is
unusable:

- **Unsupported chain:** A Risk condition does not match, and policy evaluation
continues to the next rule. This covers both a chain type that Fordefi does
not screen and a provider that returns an unsupported-chain error, for example
on an unsupported EVM chain.
- **Timeout, malformed response, or other provider error:** The result is
indeterminate, and Fordefi falls back to the default policy rule, the same
fallback used when
[simulation data is missing](/developers/simulate-transactions#override-simulation-errors).


Because an unavailable provider falls back to the default rule, we recommend
[setting a strict default policy rule](/user-guide/policies/sample-policies#set-basic-approvals).

Learn more:

- [Integrate transaction screening providers](/user-guide/transaction-screening)
- [Set policy rules on screening results](/user-guide/screening/screening-policy-rules)
- [Simulate transactions](/developers/simulate-transactions)