Skill UI

A universal infrastructure for automatically evolving and optimizing AI agents across any domain. Using LLM-driven mutation and iterative solve-observe-evolve cycles, it enhances critical components like prompts, skills, and memory by benchmarking performance against measurable datasets, enabling the creation of self-improving AI systems.

A comprehensive evaluation suite designed to benchmark the performance of Large Language Models (LLM) and Vision-Language Models (VLM) specifically for home security applications. It covers 143 diverse tests across 16 critical domains, including tool use, prompt injection resistance, contextual reasoning, alert triage, and detailed scene analysis, ensuring robustness for smart home AI systems.

This skill provides comprehensive functionality for provisioning, managing, and scaling Oracle Cloud Infrastructure (OCI) compute instances using the Python SDK. It addresses common challenges like "out of host capacity" errors by implementing robust retry logic across multiple Availability Domains (ADs). Users can select optimal Flex shapes (AMD, ARM, Intel), manage the entire instance lifecycle (start, stop, reboot, terminate), and retrieve network details, ensuring reliable cloud workload deployment.

This comprehensive runbook provides a self-service solution for managing Oracle Cloud Infrastructure (OCI) outages when official status pages are misleading or silent. It guides users through independent service health probes (Identity, Compute, Networking), automated instance recovery actions (reboot, softreset), and critical procedures for cross-Availability Domain (AD) or cross-region failover using boot volume backups to ensure business continuity.

This comprehensive skill guides users through the entire lifecycle of developing AI agent skills. From initial concept (Understand) to structured design, writing the core skill logic, rigorous testing, and final performance optimization. It ensures skills are robust, accurate, and highly triggerable across any domain, including engineering, content creation, and research.

This catalog provides a comprehensive set of foundational abstractions for building robust, scalable applications. It includes standardized interfaces for interacting with AI models, managing key-value data stores, publishing domain events, structured logging, and handling build parameters. Developers can rely on these abstractions to ensure modularity and consistency across different service layers.

This skill provides a robust, schema-based system for handling API authorization and access control. It enables developers to define detailed permission schemas (the domain layer) and create injectable, type-safe permissions abstractions (the feature layer). Use this system when implementing complex access control logic, checking granular read/write/delete permissions, or enforcing resource ownership scoping (`own` records) across API use cases.

This skill leverages the Gemini AI model to perform comprehensive and broad literature discovery across academic domains. Unlike traditional APIs, it decomposes complex research queries into sub-problems, explores topic aliases, and surfaces highly relevant papers that might otherwise be missed. It is ideal for conducting deep, multi-angled literature reviews and achieving a holistic view of a research field.

Hugging Science is a curated, LLM-friendly index specializing in AI/ML applications across diverse scientific domains, including biology, chemistry, physics, and medicine. It helps researchers discover pre-filtered datasets, state-of-the-art models, and interactive demos for solving complex scientific problems, ensuring high signal quality beyond generic search.

This assistant guides users through managing Azure resources entirely via the command line using the Azure CLI (`az`). It identifies the correct commands, prompts for missing required parameters (like resource groups or locations), constructs the full command sequence, executes the action, and diagnoses any output errors to ensure successful deployment and configuration of services across Compute, Networking, and other Azure domains.

A deep-dive session designed to stress-test a technical plan against an existing domain model. This process forces the user to resolve ambiguities, sharpen terminology, and validate decisions by cross-referencing code and existing documentation (CONTEXT.md, ADRs). It ensures architectural rigor and completeness before implementation.

An expert-level playbook detailing advanced techniques for NTLM relay and authentication coercion. Covers cross-protocol attacks targeting SMB, LDAP, HTTP, and MSSQL. Includes methods like LLMNR poisoning, PetitPotam, and advanced domain privilege escalation to bypass standard authentication controls.