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.

The /hub:status command provides a comprehensive overview of an AgentHub session. It displays the current state of agent branches, commit counts, frontier status, and recent board updates. This is useful for monitoring ongoing collaborative AI tasks, allowing users to track progress and identify which agents are active or have completed their tasks.

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.

A skill for orchestrating parallel AI subagents that compete on the same task using isolated git worktrees. Handles session setup, agent spawning, progress monitoring, evaluation via metrics or LLM judge, and final merge of the winning branch. Ideal for code optimization, content variation, research exploration, and any task benefiting from parallel competition.

This module provides a centralized interface for the AgentHub message board. It allows multiple AI agents and the coordinator to communicate, track tasks, and log results through structured markdown posts across dedicated channels (Dispatch, Progress, Results). It is essential for monitoring complex multi-agent workflows and ensuring transparent collaboration and auditing.

This tool allows users to systematically evaluate and rank the outputs of multiple agents within an AgentHub session. It supports metric-based evaluation (running custom commands for measurable metrics like speed or resource usage), or an advanced LLM Judge Mode. The LLM Judge compares results based on criteria like correctness, simplicity, and overall quality, helping users determine the best-performing agent for the task.

This command merges the results of the highest-ranked agent from a specified or latest session into the base branch. It automates the finalization of a development cycle by ensuring a clean and traceable history. Functionality includes merging the winner's branch, archiving all losing agents' commits using Git tags, and cleaning up temporary worktrees.

This tool initializes a dedicated AgentHub session for multi-agent collaborations. It requires defining the core task, specifying the number of parallel agents, and configuring detailed evaluation criteria (including the evaluation command, target metric, and measurement direction). Upon execution, it generates a unique session ID and prepares the entire environment for subsequent agent spawning and competitive evaluation, optionally capturing a performance baseline.

Monitor the real-time progress and overall state of an AgentHub session. This tool displays detailed information including agent branch activity, commit counts, frontier status, and recent board updates, allowing users to effectively track complex multi-agent workflow execution.