The dependency's LLM tool uses OPENAI_API_KEY as a fallback and sends it to Hanzo's API when the tool is invoked. This is a real cross-service credential-handling risk, but the inspected source does not show automatic collection or transmission outside that tool call, so the finding is suspicious rather than malicious.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review; source fingerprint signature matched known malicious package; routed for review
Trigger
A caller invokes the MCP `think` tool while `OPENAI_API_KEY` is set and Hanzo-specific keys are absent.
Impact
The Hanzo endpoint receives an OpenAI credential that could potentially be used against the user's OpenAI account.
Mechanism
The tool selects `OPENAI_API_KEY` as a fallback, defaults its API base to Hanzo, and attaches the selected value as a Bearer credential to a chat-completions request.
Evidence
.lpm-dependency-evidence-ec8518da-95d5-440c-ab4c-ecf007fdbd51/0/npm-linux/dist/cli.js
Network endpoints2
api.hanzo.aiAPI_URL (configurable)