@exodus/account-security

This Exodus SDK feature reports active account security warnings (GLOBAL_SCAM / LOST_PERMISSIONS) into the Safe Report.

These warnings already surface to users as modals in the desktop and mobile apps, but the Safe Report carried no record of them. When a user hits one and reaches out to support, there was previously no way to tell from the report which asset triggered the alert or which warning fired. This feature closes that gap.

Install

pnpm add @exodus/account-security

Usage

This feature is designed to be used together with @exodus/headless, where it is already wired in. See using the sdk.

It contributes a report node namespaced at accountSecurity, so its output shows up under that key whenever a Safe Report is assembled via exodus.reporting.export(). There is no exodus.* API surface, no atoms, and no plugin: the feature only reads existing state at report time.

What it reports

At report time the feature walks the account states of every enabled base asset, runs that asset's api.securityChecks({ accountState }), and collects any account that comes back insecure.

The report is null when there is no wallet or the wallet is locked. Otherwise it has the shape:

{
  summary: {
    // The most severe active warning, or null when there are none.
    warningType: 'GLOBAL_SCAM' | 'LOST_PERMISSIONS' | null
    hasWarnings: boolean
  }
  // One entry per (wallet account, asset) that failed its security check.
  warnings: Array<{
    walletAccount: string
    assetName: string
    type: 'GLOBAL_SCAM' | 'LOST_PERMISSIONS'
    reason: string
  }>
  // One entry per base asset whose securityChecks method threw.
  checkFailures: Array<{
    walletAccount: string
    assetName: string
    error: SafeError
  }>
}

Example output:

{
  "summary": { "warningType": "GLOBAL_SCAM", "hasWarnings": true },
  "warnings": [
    {
      "walletAccount": "exodus_0",
      "assetName": "tron",
      "type": "GLOBAL_SCAM",
      "reason": "Account is globally blacklisted."
    },
    {
      "walletAccount": "exodus_1",
      "assetName": "ethereum",
      "type": "LOST_PERMISSIONS",
      "reason": "Account is delegated to a non-whitelisted EIP-7702 address."
    }
  ],
  "checkFailures": []
}

Severity precedence

GLOBAL_SCAM outranks LOST_PERMISSIONS because it can block the app at startup. When both kinds are present across accounts, summary.warningType reports GLOBAL_SCAM, while warnings still lists every individual warning regardless of type.

How an asset opts in

This feature does not implement any detection logic itself, it only aggregates. The actual checks live in each asset's plugin, which exposes a synchronous api.securityChecks method:

asset.api.securityChecks({ accountState }): {
  isSecure: boolean
  type: 'GLOBAL_SCAM' | 'LOST_PERMISSIONS' | null
  reason: string | null
}

An account is only included in warnings when isSecure is false. To keep the report robust, the following are handled conservatively:

• Tokens and non-base assets: only base assets (where asset.baseAsset.name === assetName) are checked.
• Assets without securityChecks: assets that don't implement the method are ignored.
• Disabled assets: anything not enabled in enabledAssetsAtom is skipped.
• Throwing checks: if a single asset's securityChecks throws, the error is captured in checkFailures as a SafeError with the safe hint account-security: report/security-checks. The loop still continues, so one broken asset cannot sink the rest of the report.

Dependencies

The report node depends on assetsModule, accountStatesAtom, and enabledAssetsAtom, all provided by the SDK.