llm-models

Benign overall given the stated purpose of enabling multi-model LLM access through a single CLI surface. The footprint—CLI-based installation, runtime prompts routed to OpenRouter/model endpoints, and no embedded credential harvesting or unverifiable binaries—fits a legitimate developer tooling scenario. The primary security considerations are data privacy and trust in the external OpenRouter endpoints, not credential exposure or supply-chain risk. Recommend adding explicit data handling/privacy notes and ensuring users are aware prompts are sent to external services.