Skill UI

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.

Guides fine-tuning and evaluation of OpenVLA-OFT/OFT+ for robot action generation, covering LoRA adaptation, continuous action heads, FiLM conditioning, and LIBERO/ALOHA benchmarks plus deployment considerations.

A foresight agent that runs a rigorous 12-step pipeline combining Claude’s qualitative intelligence with Python’s deterministic arithmetic to analyze high-stakes future scenarios, infer reasonable horizons, and surface quantitative signals for prediction requests.

Soft Predict Future is a structured foresight engine that triggers whenever the user asks about future outcomes or scenarios. It enforces a nine-step analytical pipeline (validation, signal collection, scoring, driver extraction, etc.), highlights required searches and inputs, and directs the user to web-enabled reasoning for predictive analysis.

This tool audits the token consumption across all components (agents, skills, rules, MCP servers) within a Claude Code session. It analyzes token overhead, identifies bloat (e.g., overly long descriptions, redundant components), and generates a detailed context budget report. Use it to maintain optimal context window usage, improve performance, and prevent context saturation.

Connects to the MAXIA AI-to-AI marketplace on the Solana blockchain. This tool allows users to discover, buy, and sell AI services directly between decentralized agents. Features include comprehensive crypto intelligence, such as sentiment analysis and trending data; DeFi yield tracking; Web3 wallet security analysis (e.g., rug detection); and real-time GPU pricing, making it a comprehensive hub for Web3 AI interactions.

PydanticAI is a robust Python framework that integrates Pydantic's type-checking and validation capabilities into Large Language Models (LLMs). It allows developers to build production-ready AI agents that guarantee structured, validated outputs, moving beyond simple raw strings. Key features include supported tool calling, dependency injection for testability, and multi-model compatibility across major providers like OpenAI, Anthropic, and Gemini. Use this when reliability and correct data structure are paramount.

This skill provides guidance on implementing common architectural patterns and best practices within the PyTorch framework. It is designed for developers building sophisticated deep learning models, helping to improve structure, efficiency, and code maintainability when working with neural networks.

This skill systematically probes an AI model's core behavioral patterns across six critical dimensions, including reasoning, refusal, formatting, and grounding. It generates a comprehensive, styled HTML report with radar charts and actionable insights, allowing users to understand a model's true capabilities, compare different models, or perform compliance and red-team assessments without needing an external API key.

Handles IMO/Putnam/USAMO/AIME style problems by generating refined proofs, then running a fresh-context adversarial verifier to catch flaws, finally producing honest confidence reports and verified PDFs when possible.

An autonomous factory for generating production-ready, cross-platform AI agent skills. It accepts raw, unstructured human input—such as transcripts, PDFs, screenshots, vague ideas, or code snippets—and reverse-engineers the complete workflow logic. It bridges the gap between human intent and functional agent specifications, allowing users to automate complex processes without writing code.

AuraKit converts a single /aura command into a secure full-stack Claude pipeline with 34 modes, OWASP SEC-01~15 safeguards, runtime hooks, tiered models, token savings, and multilingual reviewers.