Skill UI

A comprehensive framework for rigorously testing and benchmarking LLM agents. It moves beyond simple pass/fail checks to assess complex behaviors, reliability metrics, and capability consistency across multiple runs. Ideal for production monitoring and identifying subtle failure modes in advanced AI agents.

Build persistent agents on Azure AI Foundry through the Microsoft Agent Framework Python SDK, hooking into AzureAIAgentsProvider for tools, hosted services, streaming, and threaded conversations.

This comprehensive guide details the architecture of memory for intelligent agents. It covers the distinction between short-term context, long-term vector storage, and advanced cognitive models (CoALA framework: semantic, episodic, procedural). Learn how to implement robust memory retrieval mechanisms, select appropriate vector databases (Pinecone, Qdrant, etc.), and utilize embedding models to build truly context-aware, persistent AI applications.

Designs and runs structured evaluation frameworks for agent systems, balancing token budgets, tool usage, and LLM scoring to compare configurations, guard production, and track regressions across complex workflows.

Framework for building stateful, multi-actor AI applications with LangGraph, covering graph construction, reducers, persistence, human-in-the-loop patterns, and ReAct agents suitable for Python production systems.

LlamaIndex is a data-centric framework for building LLM applications, offering 300+ connectors, vector indices, query engines, agents, and multimodal support to power document Q&A, chatbots, knowledge retrieval, and RAG pipelines over private data.

SGLang is a high-performance serving framework for LLMs and VLMs that uses RadixAttention prefix caching to deliver structured outputs (JSON/regex/grammar) and agentic workflows with function calls, providing up to 5× faster inference than vLLM across multi-GPU production deployments.

Use AgentFolio to discover, compare, and research autonomous AI agents, frameworks, and tools before building or integrating your own. The directory helps you map capabilities, autonomy models, and deployment patterns while collecting concrete examples for positioning workflows or vendor evaluations.

PufferLib is a high-performance library designed for advanced reinforcement learning simulations and training. It excels in parallel environment simulation, offering optimized vectorization and native support for multi-agent systems. Utilizing algorithms like PuffeRL (optimized PPO+LSTM), it achieves millions of steps per second, significantly accelerating research and deployment across single-agent and multi-agent environments.

Converts RDF context into belief-desire-intention structures so agents can track perceptions, deliberate plans, and emit explainable RDF, supporting BDI frameworks or LLMs needing structured cognition.

Documentation for the Pipecat-powered Daily framework that guides agents through building production-ready real-time voice and multimodal AI pipelines, covering low-latency audio/video/text flows, transport choices, LLM/tool orchestration, and turn/context management.

Framework for securing autonomous trading agents by sanitizing prompts, enforcing spend limits, simulating transactions, isolating wallets, and circuit-breaking before sending funds.