The reviewed credential flows are tied to the selected AI provider, including Claude credentials sent to Anthropic. The CLI's local file and command tools serve its stated agent function, and no unrelated collection or concealed execution was identified.
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
The user runs gcli or agent-cli; model requests and MCP connections occur as part of that use.
Impact
Expected CLI behavior: model services receive user prompts, while local tools perform requested project tasks.
Mechanism
The CLI sends prompts to the selected model endpoint, can use that provider's configured credentials, and exposes agent tools for local work. Claude subscription credentials are sent to Anthropic's API.
Evidence
~/.agentos/config.jsonuser-selected project files
Network endpoints8
api.deepseek.comapi.guangnao.comapi.openai.comapi.anthropic.com