Skill UI

This skill guides the end-to-end development, implementation, and validation of production-ready Python modules for Azure IoT Edge. It covers critical aspects including telemetry processing, protocol adaptation (Modbus, OPC-UA), local inference, and ensuring robust, secure edge-to-cloud messaging under variable network conditions.

This skill converts complex, narrative technical documents, research excerpts, or natural language problem descriptions into a highly structured Mathematical Problem Specification (MPS). It systematically extracts variables, constraints, objectives, and uncertainty parameters. By adhering to a strict Zero-Inference Protocol, it ensures all extracted elements are strictly grounded in the source text, making it invaluable for formalizing real-world optimization, data science, and engineering problems.

A critical discipline for managing expenses when integrating paid AI or inference APIs. This guard treats dollar cost as a third complexity dimension, mandating the definition of cost contracts (max calls per run, max budget per run, max budget per day). It enforces best practices for bounding loops, retries, and fan-out, ensuring that code bugs cannot lead to unexpected overspending. Best practice requires configuring hard caps both in the code and at the provider dashboard for defense in depth.

This meta-skill is designed to convert complex web research requests into polished, structured, and source-cited deliverables. Use it when you need a market briefing, technical report, or decision memo that requires deep web lookup and scholarly citation. It orchestrates the entire process, including preference inference, multi-source search, drafting, and optional DOCX export, ensuring the output is authoritative and actionable.

This skill manages the entire lifecycle of long-running deep learning training and inference jobs deployed on remote, ephemeral computing environments (e.g., AutoDL, RunPod, bare SSH, Slurm, K8s). It provides critical capabilities like resilient checkpointing, managing spot instance preemption, debugging resource failures (OOM, NaN), and ensuring safe billing teardown, making it essential for professional ML Ops on rented resources.

This skill provides comprehensive methods and frameworks to detect sophisticated AI attacks, including model extraction, membership inference, and model inversion, which abuse inference APIs. It focuses on monitoring per-principal query volume, input distribution, and confidence exposure. Use it for securing public or partner inference APIs, or for conducting pre-deployment red-teaming to assess the model's extractability and overall security posture against MITRE ATLAS techniques.

The `hf-mem` skill estimates the necessary VRAM or system memory required to load Safetensors or GGUF model weights for inference directly from the Hugging Face Hub. It uses efficient HTTP Range requests, avoiding local downloads. Users can determine if a model fits on their GPU or instance, including optional estimations for the KV cache and setting limits like maximum sequence length.

A comprehensive guide and skill for deploying large language models (LLMs) locally using the llama.cpp framework and GGUF quantization format. It covers searching the Hugging Face Hub, selecting optimal quantizations (e.g., Q4_K_M), running local servers, and achieving OpenAI-compatible API endpoints. Supports various hardware backends including CPU, Mac Metal, CUDA, and ROCm. Ideal for developers needing offline, private LLM inference.

Automates the creation and publication of polished, functional Gradio demos on Hugging Face Spaces for any user-provided LoRA model. This skill handles the complex workflow of designing the optimal UI, selecting the correct inference pipeline (e.g., text-to-image, video-to-video), and generating all necessary files (app.py, requirements.txt, README.md) to deploy a complete, ready-to-use AI playground.

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.

This is an advanced, structured workflow designed to conduct deep, multi-source investigations. It synthesizes current public evidence with user-provided context to generate highly structured reports. The workflow explicitly differentiates between verifiable facts, user-supplied evidence, logical inferences, and actionable recommendations, making it ideal for complex comparative analysis or continuous monitoring tasks.

This advanced skill provides a rigorous framework for assessing whether an organizational or management research manuscript is suitable for highly selective, theory-driven journals like the Academy of Management Journal (AMJ). It guides authors on necessary theoretical framing, methodological rigor (e.g., endogeneity, causal inference), structural improvements, and common pitfalls to avoid, significantly improving submission quality and maximizing acceptance chances.