@walkhi/code-relax@0.1.0-beta.45
Code Relax Windows, macOS and Linux workstation service
AI Security Review
scanned 5d ago · by lpm-firewall-aiStatic inspection confirms that global installation automatically installs a package-owned Codex skill without a separate confirmation step. This creates a bounded agent extension lifecycle risk, but the inspected source does not establish a concrete malicious payload or data transfer.
Decision evidence
public snapshot- The manifest runs a postinstall hook. It activates only when invoked directly during a global npm install, then runs this package’s setup and health check.
- Setup installs the bundled skill into a package-specific agent skills directory. The installer refuses to replace an existing same-name skill unless it is already managed by this package.
- The bundled skill guides Codex interactions about this package’s workstation service and mobile app. Its instructions say mobile message context does not add user authorization.
- The inspected account flow checks that the active ChatGPT profile matches the account identity, connects to chatgpt.com, and establishes a background messaging connection and room subscription.
- The Windows desktop helper uses a UAC elevation prompt, and the inspected skill installation is limited to the package’s own skill file.
Source & flagged code
20 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 unpkgPackage contains a possible secret pattern.
dist/src/platform/windows/rustdesk-video-probe.mjsView on unpkg · L77Findings
Affected versions and remediation
This report applies to @walkhi/code-relax@0.1.0-beta.45.
See all versions of @walkhi/code-relax for other recorded verdicts.
- Review the evidence and your use of @walkhi/code-relax@0.1.0-beta.45 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.