Skill UI

Use this command to set up a structured multi-agent collaboration environment within AgentHub. It allows you to define the core task, specify the required number of agents, and configure detailed evaluation criteria (metrics, directions, and commands) for both quantitative benchmarking and qualitative LLM judging. This is the starting point for any agent competition or complex AI workflow.

This command executes the entire AgentHub lifecycle in a single, chained invocation. It automates complex development tasks by sequentially handling initialization, capturing performance baselines, spawning multiple specialized agents, evaluating the results (via metrics or LLM judging), and finally presenting a winner for merging. Use this when running an end-to-end competition or complex refactoring task.

AgentHub spawns parallel git worktree agents that compete on tasks, evaluates their outputs via metrics or LLM judges, and merges the top branch for optimizations, research, or content variations.

The /hub:spawn command launches N subagents that work on the same task simultaneously. Each agent operates within an isolated git worktree, ensuring true parallelism and resource separation. Users can optionally specify templates (e.g., `optimizer`, `refactorer`) to guide agents toward specific strategies for improving code quality, performance, or test coverage. Ideal for complex software engineering tasks requiring diverse, competing approaches.