@metaphi-ai/hum@0.1.64
Hum: a self-improving coding agent for your terminal.
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThis is not proven credential theft or a self-dependent lifecycle chain, but it performs unconsented remote code execution during postinstall. That supply-chain behavior is a concrete unresolved risk.
Decision evidence
public snapshot- The postinstall hook automatically invokes the engine installer.
- The installer downloads and executes a remote shell or PowerShell script when uv is absent.
- The installer then installs a remotely resolved Python package into a home-directory environment during npm installation.
- The inspected Node source contains no credential upload or data-exfiltration request.
- The installer targets its own .hum engine directory and pins hum-cli to the package version by default.
- No runtime dependency on @metaphi-ai/hum is declared.
Source & flagged code
11 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to @metaphi-ai/hum@0.1.64.
See all versions of @metaphi-ai/hum for other recorded verdicts.
- Review the evidence and your use of @metaphi-ai/hum@0.1.64 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.