@tiens.nguyen/gu-cli@1.0.768
Gu CLI — the gu terminal plus the local worker that runs agent / OCR / PDF / embedding jobs on your Mac (Ollama / MLX / OpenAI-compatible).
AI Security Review
scanned 14d ago · by lpm-firewall-aiThe flagged credential flow is an interactive Gu worker login that sends the key to the configured Gu API settings endpoint, and the project MCP path requires per-folder approval before a server is runnable. The postinstall hook is limited to global PATH setup; inspected source did not establish a concrete attack.
Decision evidence
public snapshot- The postinstall hook runs install-path.mjs, which limits its PATH setup to global installs and skips local installs.
- Project MCP servers are marked runnable only when approved for that folder; approval is explicitly managed through the MCP command.
- The REPL starts first-run login only when worker configuration is missing and the user is interacting through a TTY.
- Worker-key validation sends the key to the configured Gu API settings endpoint, then saves it locally after validation.
- The inspected provider setup, proxy dispatcher, child-stop helper, worker entrypoint, and local API setup show package functions without a concrete attack chain.
- The package has a postinstall hook that may add its global binary directory to a shell profile.
- The worker key is sent to the package’s configured Gu API endpoint; source inspection shows no unrelated recipient or covert forwarding.
Source & flagged code
12 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to @tiens.nguyen/gu-cli@1.0.768.
See all versions of @tiens.nguyen/gu-cli for other recorded verdicts.
- Review the evidence and your use of @tiens.nguyen/gu-cli@1.0.768 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.