Skill UI

Real-time detection pipeline that watches entropy spikes, file system I/O, and behavioral heuristics to flag ransomware-style encryption routines before they propagate and to help trigger containment actions.

Automatically detects typosquatting in npm and PyPI by generating candidate variants from popular dependencies, querying each registry, and scoring matches with Levenshtein distance, publication recency, and download anomalies so teams can flag suspicious packages before installation.

This skill acts as a crucial bridge, automating the transition from a conceptual experiment plan to fully deployed and executable code. It parses structured experiment roadmaps, implements necessary training/evaluation scripts, manages dependencies, and deploys the experiment setup to GPU environments. It ensures reproducibility through mandatory code review and structured result collection.

A library-agnostic Flutter/Dart code review checklist that walks through widget best practices, state management patterns, Dart idioms, performance, accessibility, security, and clean architecture, helping reviewers spot regressions in any app stack.

Guides researchers through organizing assumptions, invariant objects, notation, and derivation maps to turn loose goals, messy notes, or partial formula chains into a coherent package suitable for paper-style theory drafts or blocker reports.

This skill guides the transformation of validated research ideas into a comprehensive, reviewer-ready grant proposal. It supports major international and national funding bodies (e.g., NSF, KAKENHI, ERC, NSFC) by handling specific requirements like narrative arc design, budget justification, and adherence to agency-specific formatting. It structures the process into a multi-stage pipeline: discovery, novelty check, drafting, review, and revision.

Provides pure-function CDSS modules for EMR integration, covering drug interaction checks, dose validation, NEWS2 scoring, and alert classification to keep patient-safety-critical workflows testable and reliable.

Guidelines for EMR/EHR systems that prioritize patient safety, clinical flows, medication oversight, CDSS integration, accessibility, and audit-ready encounters for healthcare apps.

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.

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.

Detects DCOM lateral movement by triangulating Sysmon Event ID 1/3, Windows security logs, and RPC traffic patterns tied to MMC20, ShellWindows, and ShellBrowserWindow abuse, enabling SOCs to hunt and tune detections across AD domains.