Skill UI

Guides developers through scikit-learn workflows for classical supervised and unsupervised tasks, data preprocessing, evaluation schemes, hyperparameter tuning, and end-to-end pipelines when building interpretable production models.

This comprehensive guide outlines a rigorous, human-in-the-loop engineering workflow for leveraging advanced AI agents to complete complex implementation tasks. It mandates defining clear completion criteria, decomposing work into manageable units, and strategically routing models (Haiku, Sonnet, Opus) based on task complexity. The workflow emphasizes continuous evaluation, regression testing, security review of invariants and edge cases, and strict cost discipline to ensure robust and reliable software development using AI assistance.

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 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.

A lightweight CLI tool for systematically comparing various coding AI agents (such as Claude Code, Aider, and Codex) head-to-head. It measures crucial metrics—pass rate, API cost, execution time, and consistency—on reproducible tasks defined via YAML manifests. Use this to objectively select the best agent for your codebase or to validate model updates before production deployment.

Applies two-axis scoring, mandatory devil’s advocate reasoning, and persistent history to keep post-task self-reviews honest, resistant to inflation, and suitable for Claude Code sessions.

This skill guides users on setting up and running LLM-as-judge evaluations using Arize. It covers defining evaluators (including prompts, classification choices, and models) and executing tasks against real data spans or experiments. It supports key metrics like hallucination, faithfulness, and correctness, and facilitates continuous monitoring for comprehensive LLM performance analysis.

A comprehensive resource detailing the use of scikit-learn, the industry-standard Python library for classical machine learning. Use this skill for the entire ML lifecycle, including supervised tasks (classification, regression), unsupervised tasks (clustering, dimensionality reduction), data preprocessing, model evaluation, and building production-ready pipelines.

This skill performs a comprehensive usability audit and heuristic evaluation of interactive user interfaces (UIs). It goes beyond mere visual aesthetics, assessing whether a system is intuitive, easy to learn, and functional for a first-time user. Use it when auditing complex tools, editors, or apps to ensure the user can complete primary tasks without needing extensive instructions or guidance.

This skill executes the DeepSWE benchmark, a comprehensive coding-agent evaluation suite. It uses OpenRouter and mini-swe-agent to score various Large Language Models (LLMs) on 113 tasks, providing reproducible results for model comparison, development assessment, and research benchmarking. Use this when a standardized, deep evaluation of LLM coding capabilities is required.

A lightweight CLI tool for systematically comparing various coding AI agents (such as Claude Code, Aider, and Codex) head-to-head. It measures crucial metrics—pass rate, API cost, execution time, and consistency—on reproducible tasks defined via YAML manifests. Use this to objectively select the best agent for your codebase or to validate model updates before production deployment.

This framework guides AI agents to critically assess their own output across five key axes: Accuracy, Completeness, Clarity, Actionability, and Conciseness. It provides a structured, evidence-based scorecard and concrete suggestions for improvement, turning every task completion into a deep learning opportunity.