Skill UI

This tool automates the generation of high-quality, publication-ready academic illustrations and diagrams. It employs a sophisticated, multi-stage, iterative workflow supervised by Claude, covering prompt planning, layout optimization, style verification (CVPR/NeurIPS standards), image rendering, and final strict visual review. Use it when generating complex figures for research papers.

Correlates OSINT findings across usernames, emails, social media, domains, breach records, and dark-web mentions using AI/LLM reasoning to build unified target profiles with confidence scoring and link analysis.

This skill guides the entire process of writing a research paper rebuttal following submission to a major conference. It parses complex, multi-reviewer feedback, structures responses according to strict venue rules (e.g., text-only, per-reviewer threads), and ensures all claims are grounded in the original paper, preventing fabrication or overpromising. It manages follow-up rounds and maintains academic rigor.

This tool assesses whether coding agents adhere to defined skills, rules, or agent definitions. It automatically generates scenarios with varying prompt strictness, executes the agent, and generates a comprehensive report detailing the compliance rate of tool calls against predefined workflow specifications using advanced LLM classification.

Tests and validates ransomware recovery procedures by exercising backup restores, measuring RTO/RPO, checking sequencing, and verifying clean restores plus post-restore security hardening to ensure resilience.

Run Hugging Face models directly in JavaScript/TypeScript for text, vision, and speech tasks without backend servers, using Transformers.js in Node.js or modern browsers with WebGPU, quantization, and pipeline abstractions for efficient inference.

This skill automates the process of renting, configuring, and managing GPU instances on vast.ai for deep learning and AI tasks. It analyzes the task requirements (VRAM, GPU count, estimated time) from the experiment plan or user input, searches for the most cost-effective and reliable GPU offers, and manages the entire resource lifecycle, ensuring efficient and economic AI training.

Assesses AI/ML systems for prompt injection, jailbreak, model inversion, data poisoning, and agent tool abuse using MITRE ATLAS mapping, signature detection, and scoring metrics before deployment.

Provides an expert methodology to significantly reduce AI API spending (40-80%) without sacrificing user quality. The guide covers cost auditing, optimizing existing systems, and designing cost-efficient architectures using techniques like model routing, prompt caching, output length control, and prompt compression.

Helps teams treat prompts like production code by providing registry, eval, and governance workflows for versioning, testing, and promoting prompts safely across environments.

This harness transforms standard code environments into fully persistent, autonomous agent systems. It leverages native features like scheduled tasks (Crons), remote dispatching, and structured memory storage (MCP) to enable continuous, self-directing automation. Ideal for building AI assistants that require long-term context, recurring workflows, and complex, multi-step operations, eliminating the need for external agent frameworks like AutoGPT.

Executes training, fine-tuning, inference, and batch jobs on Modal’s zero-config GPU cloud so you just write Python, run `modal run`, and let Modal auto scale, manage secrets, and halt billing when finishes.