Skill UI

Diagnoses full-stack failure modes in agent and LLM applications, including wrapper regressions, memory contamination, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings and code-first fixes for production agent systems.

A lightweight utility designed for fetching content from a single URL via HTTP or HTTPS. It supports various methods (GET, POST, PUT, DELETE) and allows configuration of timeouts and maximum response body sizes. This tool is ideal for quick, simple data extraction steps within a workflow DAG, offering significant performance gains over complex LLM agent loops. Note that complex tasks like JS rendering or multi-page crawling require dedicated sub-agents.

An open-source automation framework for conducting multi-turn, adversarial red-teaming against conversational LLMs. PyRIT simulates real-world attacker behavior using advanced orchestrators (Crescendo, TAP) and scorer feedback loops. It is designed for comprehensive security assessments of AI agents and chatbots, going far beyond single-shot prompt testing to evaluate robustness under gradual or adaptive attacks.

Diagnoses full-stack failure modes in agent and LLM applications, including wrapper regressions, memory contamination, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings and code-first fixes for production agent systems.

A comprehensive guide covering patterns and architectures for automating code execution using LLMs. It spans simple sequential pipelines (e.g., daily dev scripts) to advanced, RFC-driven multi-agent Directed Acyclic Graph (DAG) orchestration. Ideal for building CI/CD style continuous development pipelines, managing complex state, and automating quality gates without human intervention.

Provides comprehensive configuration management for local LLM inference using RuVLLM. Users can select models, perform task-specific fine-tuning with MicroLoRA adapters, or implement continuous, real-time adaptation using SONA. This tool guides the entire advanced LLM workflow, covering status checks, configuration generation, and advanced adaptation loops.

Architecture and integration sketches for LLM applications, covering RAG retrieval, tool dispatch, agent loops, caching, privacy boundaries and verification requirements before production use.

Comprehensive guide to LLM Operations (LLMOps). Covers advanced topics like Retrieval-Augmented Generation (RAG), vector database management (Pinecone, Chroma), embeddings, fine-tuning, prompt engineering, and cost optimization. Essential for bridging the gap between AI prototypes and reliable, scalable, production-grade AI systems.

This skill provides methods for reliably extracting typed, validated, and structured data from Large Language Models (LLMs), moving beyond simple text parsing. It covers best practices across major APIs (OpenAI's response_format, Anthropic's tool_use, Google's responseSchema) and local models. Key workflows include defining schemas (Pydantic/Zod), implementing constrained decoding, and building robust retry loops for production-grade data pipelines.

Design and operate production LLMOps platforms with CI/CD, model promotion workflows, evaluation gates, rollback, and governance for cloud and self-hosted inference.