Skill UI

Automated harness for healthcare deployments that runs Jest suites across CDSS logic, PHI exposure, data integrity, clinical workflow, and integration gates, aborting releases on any critical failure to keep patient safety non-negotiable.

Patterns for protecting patient and clinician data across healthcare applications, covering classification, row-level access control, auditing, and deployment checklists to meet HIPAA, DISHA, GDPR, and other regulations.

This skill is designed for performing local, on-premise evaluations and benchmarking of models hosted on the Hugging Face Hub. It facilitates running specialized evaluation frameworks like `inspect-ai` and `lighteval` against local hardware. Users can select optimal inference backends, including `vllm` for high throughput, or fall back to Hugging Face Transformers or `accelerate`, ensuring comprehensive smoke testing and task selection for LLMs. It focuses purely on the local execution process, separate from remote jobs or publishing workflows.

Gradio is a powerful Python library designed for building and sharing machine learning demo applications and user interfaces (UIs). It allows developers to quickly create interactive web prototypes, from simple interfaces to complex chat interfaces, streamlining the process of demonstrating ML models without needing extensive front-end coding knowledge. It supports advanced features like custom layouts and streaming outputs.

Trackio is a comprehensive library for logging, visualizing, and managing ML experiment metrics and alerts. It allows developers to use a Python API to log metrics and set diagnostic alerts (INFO/WARN/ERROR) during training. Furthermore, a dedicated CLI enables metric retrieval and dashboard viewing, syncing data to Hugging Face Spaces for persistent, remote monitoring, making it ideal for autonomous ML agent workflows.

Implements remote browser isolation with URL risk scoring, isolation policies, content disarming/reconstruction, DLP controls, and Secure Web Gateway plus ZTNA integration to harden web browsing against exploits and credential theft.

Sets up Microsoft Purview data loss prevention policies to protect sensitive content across Exchange, SharePoint, OneDrive, Teams, Power BI, and managed endpoints by defining sensitivity label taxonomies, custom regex types, endpoint controls, and monitoring with Activity Explorer plus PowerShell/Graph automation.

Automatically transforms a full academic research paper into a structured, narrative conference presentation deck. It adapts the content depth, slide count, and structure based on specified talk types (e.g., spotlight, oral) and target conferences. Outputs include Beamer LaTeX source, PDF slides, and editable PPTX with speaker notes, ensuring a coherent narrative flow for academic talks.

Guides analysts through using binwalk to scan, extract, mount, and inspect firmware images, covering entropy analysis, nested archive handling, filesystem mounts, and string hunting for credentials or config data during IoT/security research.

This gate systematically evaluates experimental outcomes against intended research claims. It collects metrics from various sources (W&B, logs, etc.) and uses an objective judgment model to determine if the data supports, partially supports, or fails to support the hypothesis. It provides specific evidence gaps and suggests concrete next steps, guiding the research workflow before writing papers or finalizing designs.

Safety Guard prevents destructive operations on production systems or during autonomous agent runs by intercepting risky commands, restricting write access to specified directories, and combining both protections when needed.

Guides you through profiling scripts, processes, GPUs, memory, or interconnects with external tools and custom instrumentation, then summarizes CPU, memory, and communication bottlenecks plus actionable recommendations.