@embarkai/ui-kit@0.4.40
React UI components and hooks for EmbarkAI authentication and Account Abstraction
AI Security Review
scanned 4mo ago · by lpm-firewall-aiStatic inspection found high-risk browser primitives, but they are consistent with the package's declared auth/account-abstraction UI purpose and are user/runtime invoked. Scanner claims of remote asset decode-execute and child_process were not substantiated in the inspected source.
Decision evidence
public snapshot- dist/iframe/main.js ships dkls23_wasm_bg.wasm and wasm-bindgen glue with new Function.
- dist/index.cjs creates a hidden auth iframe and exchanges postMessage requests for auth/signing flows.
- dist/index.cjs and dist/iframe/main.js read/write localStorage keyshares and auth/session metadata when user invokes wallet backup/auth flows.
- package.json has no preinstall/install/postinstall lifecycle hooks and exports only dist entrypoints/styles.
- No child_process, shell execution, destructive filesystem writes, or AI-agent control-surface writes found in package files.
- Network calls are aligned with auth, TSS, vault, KYC, OAuth, price, and cloud-backup product features.
- dist/iframe/main.js fetches local dkls23_wasm_bg.wasm via import.meta.url; no remote asset decode-and-execute path was confirmed.
- postMessage handlers check expected iframe origin or SDK message structure/HMAC before sensitive iframe actions.
Source & flagged code
13 flaggedSource fetches a remote non-code asset, decodes its contents, and dynamically executes the decoded payload.
dist/iframe/main.jsView on unpkg · L13Package contains a possible secret pattern.
dist/iframe/main.jsFindings
Affected versions and remediation
This report applies to @embarkai/ui-kit@0.4.40.
See all versions of @embarkai/ui-kit for other recorded verdicts.
- Review the evidence and your use of @embarkai/ui-kit@0.4.40 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.