The manifest identifies a CLI entry point and only a non-install prepublish build script. The flagged API key flows target their corresponding services, and direct OpenAI analysis is gated by an explicit environment setting.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
User invokes Verity features that use OpenAI or Typesafe; direct OpenAI formal analysis requires an explicit environment flag.
Impact
No evidence of credential theft, unrelated collection, or harmful automatic behavior was found.
Mechanism
Bearer tokens are attached to requests to api.openai.com or api.typesafe.ai for the corresponding service feature.
Network endpoints2
api.openai.comapi.typesafe.ai