Skill UI

This skill is designed for running comprehensive evaluations of Hugging Face Hub models on local hardware. It supports multiple evaluation frameworks, including `inspect-ai` and `lighteval`, and provides flexibility in backend selection (e.g., `vllm`, Transformers, or `accelerate`) to optimize for performance and compatibility. Use this for local GPU benchmarking, smoke testing, and backend selection, but note that it does not handle remote job orchestration or result publishing.

Transforms the AI into a highly personalized virtual tutor, adopting a defined persona, teaching style, and emotional relationship. This skill uses the Socratic method and advanced psychological principles (like Self-Determination Theory and memory scaffolding) to guide users through discovery learning. It doesn't just deliver answers; it fosters genuine engagement, maintains continuity across sessions, and motivates the user to learn by simulating the experience of a caring, challenging human mentor.

This skill is designed to translate vague, colloquial, or descriptive biological details (such as behaviors, appearances, or cultural uses) into precise scientific or common species names. It is essential for constructing effective search queries in biological databases when the question does not use formal taxonomic names. By accurately identifying candidate species names from descriptive clues, users can drastically improve search relevance and accuracy.

Detects goal hijacking vulnerabilities where malicious instructions are hidden within external content sources processed by the agent, such as retrieved documents (RAG), uploaded files, or web page content. This skill simulates these attack vectors to ensure the agent only uses the provided context for the stated task and does not execute embedded hidden commands.

This skill provides the frontend interface for image generation tasks. It has been moved to a dedicated GitHub repository containing other image-generation skills. It runs only when explicitly called by name and does not invoke models directly. Users can install it via the npx skills command to access the web interface for generating images.

This skill guides the architectural design of an Itô Data Atlas agent for advanced quantitative research. It facilitates market discovery, constructs candidate prediction-market baskets, drafts necessary parameter changes, and integrates human-in-the-loop review. It is strictly for workflow and architectural planning, and does not execute live trading orders.

This guide instructs authors on how to package theoretical proofs for submission to COLT (Conference on Learning Theory). Since COLT does not have a formal artifact track, it emphasizes treating the proof appendix as the primary 'artifact.' It details best practices for creating a verifiable proof ledger, rigorously tracking constants and external results, and managing supplementary code to ensure maximum academic credibility.

A sophisticated framework designed to simulate complex legal interactions across various domains, including tenancy law, business contracts, startup disputes, and employment claims. This tool is ideal for demonstrating AI-powered legal triage systems, training personnel on responsible AI use, or creating realistic proof-of-concepts for legal technology presentations. Note: This skill is strictly for educational and demonstration purposes and does not provide legal advice.

This tool serves as a model-routing advisor for high-stakes legal translation tasks, covering contracts, statutes, and case law. It does not perform the translation itself, but instead provides expert decision support by recommending the optimal primary and fallback LLMs based on language pair, required privacy level, cost, and speed. Crucially, it emphasizes that human legal-linguist review is mandatory, as no single LLM can guarantee certified legal accuracy.

This skill guides users on running evaluations against models on the Hugging Face Hub using local hardware. It covers using inspect-ai and lighteval frameworks with local inference backends like vllm, Transformers, or accelerate. It includes task selection, smoke testing, and hardware guidance for different model sizes. It does not cover remote orchestration or publishing results to the community workflow.

This skill enables users to generate, compose, and edit product images, brand visuals, and social graphics using the official AI Image Generator package on the Beatra service. It requires explicit approval for paid hosted work and integrates with tools like Claude Code. The package is pinned by digest for security and does not self-update.

This skill fetches public webpages and converts them into clean Markdown format suitable for AI agents. It allows users to read, summarize, quote, or cite webpage content while preserving the original source URL. The workflow involves sending a GET request to a specific endpoint to retrieve the rendered content. It is designed for read-only access and does not support private pages or content modification.