Skill UI

A comprehensive offensive checklist for assessing the security and robustness of AI and Large Language Model (LLM) applications. It covers advanced adversarial techniques such as prompt injection, jailbreaking, model extraction, data poisoning, and analyzing system vulnerabilities across components. Essential for red-teaming and security assessment.

This skill translates complex legal jargon (legalese) from contracts and agreements into clear, plain English at an 8th-grade reading level. It systematically identifies and rewrites material clauses, replacing archaic terms with everyday equivalents. Crucially, it also flags potentially misleading or high-risk sections, providing a clear, non-technical 'What You Are Actually Agreeing To' summary for non-lawyers.

This skill provides implementation patterns for utilizing Runway's AI video generation and creative platform. It covers setting up the SDK, generating videos from images, and handling common API errors (like insufficient credits or unauthorized access). It is designed for developers needing to integrate secure, high-quality AI video workflows into their applications.

This comprehensive benchmark suite evaluates the performance of Vision-Language Models (VLM) on complex video anomaly detection tasks using the SmartHome-Bench dataset. It covers a wide range of real-world smart home scenarios, including security breaches, elderly fall detection, and unusual pet behavior. Unlike single-frame analysis, this benchmark requires deep multi-frame video understanding, making it a robust tool for testing the advanced AI capabilities of video monitoring systems.

This comprehensive checklist guides the operational setup (MLOps) for AI applications using Together AI's OpenAI-compatible API. It covers critical production concerns, including secure API key management, implementing robust rate limit handling, building error resilience features (circuit breakers, exponential backoff), and setting up comprehensive monitoring for latency, cost control, and model availability. Essential for reliable, scalable LLM deployment.

Offensive cloud security methodology for AWS, Azure, and GCP. Covers identity discovery, credential harvesting, enumeration, privilege escalation, persistence, data exfiltration, lateral movement, serverless abuse, Kubernetes paths, and detection evasion during authorized assessments.

A comprehensive guide detailing advanced security threats in Machine Learning and AI systems. Covers model supply chain attacks (RCE via pickle, poisoning), adversarial examples (FGSM, PGD, physical-world), data privacy breaches, and LLM-specific vulnerabilities. Essential for building robust, secure AI pipelines.

A comprehensive expert playbook detailing advanced Large Language Model (LLM) prompt injection techniques. Covers direct methods (instruction override, role play), indirect attacks (RAG poisoning, web data manipulation), and sophisticated tool/function calling abuse for data exfiltration and security bypass. Essential for rigorous AI application security testing.

This skill automatically enriches research wiki pages that are left as bare scaffolds after initial ingestion. It reads multiple external sources (like arXiv/alphaxiv) and fills in 10 structured `_TODO._` sections (e.g., Problem/Gap, Key Results, Open Questions) with detailed prose summaries. Use this skill to transform incomplete research scaffolds into a comprehensive, structured knowledge base.

This skill uses Python and boto3 to analyze AWS CloudTrail logs. It establishes statistical baselines of normal API activity and detects various anomalies, such as unusual event sources, first-time API usage, geographic IP deviations, and high-frequency calls. This is critical for identifying credential compromise, privilege escalation, and unauthorized resource access in a security incident.

Automated system designed to monitor paste sites (e.g., Pastebin, GitHub Gists) for leaked credentials, API keys, and sensitive data dumps. It uses advanced scraping, keyword matching, and regex patterns to provide early detection of data breaches and security incidents.

A comprehensive security tool designed to detect and defend against subtle prompt injection attacks hidden within untrusted multi-modal content, such as web pages, PDFs, and images. It scans for obfuscated payloads (e.g., hidden CSS, metadata, zero-width characters) by combining content normalization, heuristic analysis, and dedicated ML detectors, ensuring that LLM agents only consume sanitized, trustworthy input.