Skill UI

This skill provides a detailed, side-by-side comparison of advanced LLM agent evaluation frameworks, such as ruflo and HAL. It helps researchers and developers identify capability gaps—including real code execution, full web browsing, and structured file parsing—and prioritize architectural improvements for robust agent benchmarking and performance tracking.

Guides users through the entire end-to-end process of running the GAIA benchmark. It covers environment validation, cost estimation, model execution, and generating a signed, HAL-compatible submission package for leaderboards. This ensures high integrity and reproducibility for model performance evaluation.

This tool provides deep inspection and auditing capabilities for GEPA (genetic-evolution prompt-adaptation) genomes by utilizing the internal `@metaharness/darwin/gepa` library. It is crucial for advanced AI development pipelines, allowing users to validate the structural integrity of genomes, render the full system prompt derived from a genome, or classify specific failure modes found in an execution transcript. Ideal for CI/CD gating and debugging complex LLM applications.

A powerful tool designed to scaffold sophisticated, custom AI agent harnesses using the metaharness CLI. It supports various templates (e.g., coding, devops, marketing) and AI hosts (e.g., Claude, Codex). Crucially, it operates with maximum safety, defaulting to a dry-run mode and refusing to write to the calling repository root, ensuring controlled and predictable development workflows.

This tool executes the advanced `metaharness/darwin` security benchmark (ADR-155). It rigorously evaluates the security detection capabilities of AI systems against a controlled ground-truth corpus, providing critical metrics like TPR, FPR, and patch-pass rates. It compares the champion model against multiple baselines, offering empirical data crucial for iterative security improvement and drift detection.

This skill provides comprehensive security and monitoring for Large Language Model (LLM) powered trading agents. It helps to audit trade decisions, ensure compliance with risk parameters, and mitigate potential risks associated with automated financial operations.

A state-of-the-art memory search engine providing multi-faceted retrieval strategies. It supports default dense semantic matching, advanced hybrid search (sparse + dense with RRF), deep multi-hop Graph RAG for complex reasoning, and 'Smart' search utilizing query expansion, recency boosting, and MMR diversity reranking. Ideal for retrieving nuanced, diverse, and time-sensitive information from large, complex knowledge bases across multiple namespaces.

Pilot Protocol is an open-source overlay network providing AI agents with first-class network citizenship. It offers permanent virtual addresses, encrypted P2P communication, NAT traversal, and a structured app store. Use this skill when agents require stable networking across cloud changes, need direct, secure messaging without shared accounts, or must integrate structured external data services (e.g., live crypto prices) via local, typed capabilities.

This tool is designed to help optimize and refine prompts used with Large Language Models (LLMs). By structuring inputs, it ensures that the instructions given to AI systems are clear, comprehensive, and contextually rich, resulting in significantly higher quality and relevance of the generated output. Useful for developers, writers, and researchers.

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.

A structured framework for managing complex, iterative, and high-stakes decision processes (e.g., stochastic optimization, ensemble comparison, model rollouts). It mandates the creation of a detailed, auditable 'ledger' that records every prior result, new input, search space metric, and decision mark. This tool enforces rigorous evidence trails and defaults to safe, dry-run modes, ensuring that confidence is built through verifiable, step-by-step proof rather than assumed certainty.

RouterBase acts as a unified, OpenAI-compatible API gateway, allowing applications to route requests across diverse large language models (LLMs) like GPT, Claude, and Gemini. It helps developers migrate existing OpenAI SDK calls, implement sophisticated primary/fallback model selection based on cost, latency, and quality, and manage multimodal payloads (image, audio, video) from a single, standardized endpoint.