eve@0.74.0
Filesystem-first framework for durable backend AI agents that run anywhere.
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe source behind the high-risk hints shows package-aligned Vercel authentication and ordinary CLI, MCP, and channel-authentication behavior. The manifest has no install lifecycle hook, and the inspected code establishes no concrete attack surface.
Decision evidence
public snapshot- The manifest identifies the CLI entrypoint and ends with build, development, and test scripts; it contains no install lifecycle hook.
- The CLI starts child processes to run the package’s own build steps when its compiled entrypoint needs rebuilding.
- Vercel OAuth endpoints are restricted to Vercel origins, and the Vercel access check sends a bearer token to Vercel’s AI Gateway.
- The MCP skills module registers handlers to list and read supplied skill resources.
- HTTP Basic authentication compares the supplied credentials with the configured strategy; the public channel verifier delegates to that check.
Source & flagged code
9 flagged · loading sourceSource spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.
bin/eve.jsView on unpkg · L2Package source references dynamic require/import behavior.
bin/eve.jsFindings
Affected versions and remediation
This report applies to eve@0.74.0.
See all versions of eve for other recorded verdicts.
- Review the evidence and your use of eve@0.74.0 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.