The flagged Xiaomi host appears only in an error example for an operator-configured model provider, not as a default credential destination. The reviewed CLI and console behaviors are consistent with the package’s stated purpose, and no malicious automatic behavior was found.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review
Trigger
Invoking the CLI or explicitly running a model-backed feature.
Impact
No unsolicited credential collection, unrelated data transfer, or harmful default action was found in the reviewed paths.
Mechanism
The CLI dispatches commands; the console launcher uses loopback, and model requests use the configured provider endpoint or the OpenAI default.
Evidence
package.jsondist/index.jsdist/commands/health-codex.jsdist/commands/console.jsdist/services/console-launcher.jsdist/commands/runtime-internalization-run-once.js.lpm-dependency-evidence-d0c98692-019f-4737-896a-b91669f068bb/0/npm-linux/dist/runtime/pain-signal-runtime-factory.jsdist/services/quality-scorecard/strong-model-gate.js