eve@0.52.3
Filesystem-first framework for durable backend AI agents that run anywhere.
AI Security Review
scanned 1mo ago · by lpm-firewall-aiNo concrete malicious install hook, persistence, broad control-surface mutation, or unconsented exfiltration was established. Downgrade to a warning because opaque bundled code and the runtime authentication/network path prevent a grounded clean decision.
Decision evidence
public snapshot- A historical release of the same package is present under an alias, which warrants supply-chain scrutiny.
- The user-invoked CLI can spawn local build commands if its compiled entrypoint is absent.
- Several very large minified vendored bundles limit complete static behavioral coverage.
- The bundled OIDC helper includes the host name in a request header and contacts Vercel during token refresh.
- package.json has no preinstall, install, postinstall, or prepare hook.
- The historical package alias is in devDependencies, not runtime dependencies.
- The CLI command execution is reached through direct execution of the eve bin.
- The observed OIDC endpoint is Vercel-aligned and the code is token-refresh functionality; no unconsented credential exfiltration was confirmed.
Source & flagged code
17 flagged · loading sourcePackage contains a critical-looking secret pattern.
dist/src/compiled/@vercel/blob/index.jsView on unpkg · L28RSA private key in dist/src/compiled/@vercel/blob/index.js
dist/src/compiled/@vercel/blob/index.jsView on unpkg · L28Findings
Affected versions and remediation
This report applies to eve@0.52.3.
See all versions of eve for other recorded verdicts.
- Review the evidence and your use of eve@0.52.3 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.