Skill UI

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.

This skill provides access to the Anthropic Claude Managed Agent runtime, which is a cloud-native, persistent environment for running complex, long-running, or asynchronous AI tasks. Unlike local WASM agents, the managed runtime provides a full container, persistent filesystem, and robust state management across multiple turns. It supports the entire agent lifecycle: creation, prompting, status inspection, event logging, and mandatory termination.

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.

A mechanism for spawning hierarchical, nested sub-agents up to five levels deep, designed specifically for deep task delegation while managing context limitations. Instead of parallel execution, this ensures that each layer of decomposition receives a fresh context window, preventing the parent agent's context budget from being overwhelmed by complex, multi-step workflows. Ideal for research, verification, and synthesis processes.

This tool manages the lifecycle of learned intelligence patterns, implementing the crucial DISTILL and CONSOLIDATE phases of the 4-step intelligence pipeline. It enables the capture of successful task knowledge, supports multi-domain adaptation via SONA or MicroLoRA, and consolidates learned patterns into long-term storage to prevent catastrophic forgetting.

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.

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.

This tool manages RVF (Ruflo Vector Format) files, providing comprehensive functions for handling, transferring, and backing up agent memories. It supports listing, exporting, importing, migrating, and generating statistics for complex vector embeddings and associated metadata, ensuring knowledge portability across different projects and platforms.

This tool assesses whether coding agents adhere to defined skills, rules, or agent definitions. It automatically generates scenarios with varying prompt strictness, executes the agent, and generates a comprehensive report detailing the compliance rate of tool calls against predefined workflow specifications using advanced LLM classification.

This skill intercepts the AI response flow to allow users to explicitly control the desired depth, length, and token usage of the generated answer before the model begins writing. By analyzing the prompt's complexity and the user's intent (e.g., 'short version', 'detailed answer'), it estimates the appropriate token range and prompts the user to select a granularity level (25%, 50%, 75%, or 100%).