@metaphi-ai/hum@0.1.63
Hum: a self-improving coding agent for your terminal.
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThis is an unconsented install-time remote-code bootstrap, but the observed filesystem target is the package's own engine directory and the inspected JavaScript contains no concrete credential theft or broad agent-control mutation. It warrants a warning rather than a block under the stated lifecycle policy.
Decision evidence
public snapshot- The package registers an automatic postinstall hook.
- The postinstall hook imports and invokes the engine installer when the bundled engine version is absent.
- The installer downloads and executes Astral's uv installer through a shell, then installs a Python engine package.
- The CLI can later launch detached background updates that execute a remotely fetched installer script.
- The inspected JavaScript has no browser credential collection or secret exfiltration logic.
- The frontend network socket is restricted to 127.0.0.1 for its local engine.
- The installation target is the package's own .hum directory rather than a broad agent configuration surface.
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@0.1.63.
See all versions of @metaphi-ai/hum for other recorded verdicts.
- Review the evidence and your use of @metaphi-ai/hum@0.1.63 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.