Show HN: OpenCode Agent Router – Jev Controlled Subagents (github.com)

A new OpenCode plugin, the OpenCode Agent Router, has been introduced, enabling the delegation of tasks to subagents based on a model selection process guided by a component named Jev. This plugin enhances OpenCode v2, allowing it to route each new task to one of the specified AI models—such as OpenAI's GPT-5.4, Anthropic's Claude Sonnet 4.6, or Google's Gemini 3 Pro—before creating a child session. Users can set a confidence threshold for Jev’s recommendations and configure the plugin through specified JSON settings, streamlining task management by automatically selecting the most appropriate model for execution. This development is significant for the AI/ML community as it optimizes task execution by leveraging multiple models, enabling developers to utilize specific expertise for varied tasks. The flexibility to adjust parameters like confidence threshold and to incorporate agent IDs fosters a more dynamic and responsive AI environment. Additionally, the plugin caters to the complexity of modern AI workflows, making it easier for users to manage and route tasks efficiently across different models, while preserving the integrity of resumed sessions and explicit model calls.