Skill UI

This skill executes an autonomous, iterative research review loop, simulating a rigorous peer-review process. It repeatedly uses an external OpenAI-compatible LLM via API calls to score, critique, and identify weaknesses in a given research topic or document. The process continues to refine the content by implementing suggested fixes until the LLM provides a positive assessment or the maximum round limit is reached. Ideal for academic writing, technical documentation, and refining complex proposals.

This skill automates the iterative process of academic paper refinement, simulating the rigorous peer-review cycle of top-tier conferences (e.g., NeurIPS, ICML). It connects with an external LLM (MiniMax) to receive multi-round, critical feedback, tracks state persistence, and guides the user through necessary revisions until the research is deemed submission-ready. Ideal for maximizing the quality of technical writing.

Uses Sysmon telemetry to detect MITRE T1055 process injection techniques by monitoring CreateRemoteThread, ProcessAccess, DLL loading, and ProcessTampering events, correlating them with process creation and known-good filters for threat hunting.

This advanced skill systematically verifies the novelty of a proposed research method or idea against the latest academic literature. It guides the user to extract core technical claims, performs multi-source searches across arXiv, Google Scholar, and major conferences, and utilizes advanced cross-model reasoning to generate a structured Novelty Report detailing overlap, suggested positioning, and overall risk assessment. Essential for academic researchers.

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.

This mode enables ultra-compressed communication, drastically reducing token usage (up to 75%) while retaining full technical accuracy. It simulates speaking in a highly terse, telegraphic style. Users can select from multiple intensity levels (lite, full, ultra, and classical Chinese variants) to tailor the level of abbreviation needed for specific contexts, making it ideal for API interactions or constrained API calls.

Creates high-fidelity visuals through structured JSON prompts and a Python generator script, handling subjects, styles, references, and technical specs for characters, scenes, or products.

This skill creates high-quality videos from structured JSON prompts and optional reference images, then runs the prescribed Python script to produce MP4 output tailored to the requested scene, style, and technical specs.

This utility compresses natural language documents (notes, preferences, memory files) into a highly condensed, 'caveman-speak' format. Its primary function is drastically reducing input tokens for efficiency while meticulously preserving all technical substance, including code blocks, URLs, file paths, and markdown structure. Ideal for optimizing memory consumption when interacting with AI models.

Provides patterns and techniques for building robust evaluation loops for AI agents. This skill guides the implementation of self-critique, evaluator-optimizer pipelines, and test-driven refinement workflows. Use it to significantly enhance the accuracy and quality of complex outputs like code, reports, and analyses by implementing structured, iterative improvement cycles (Generate → Evaluate → Critique → Refine).

Provides comprehensive patterns and techniques for securing AI agent systems. Learn how to enforce governance policies, including tool allow/block lists, content filtering, rate limiting, and human approval workflows. Essential for building robust, compliant, and trustworthy agents in multi-agent or production environments that interact with external APIs and sensitive data.

This guide outlines the structure and content guidelines for an AGENTS.md file. This file serves as a comprehensive 'README for agents,' providing detailed, technical context and operational instructions (such as setup, development workflow, testing procedures, and code style) that enable advanced AI coding agents to interact with and understand a codebase effectively. It greatly complements traditional READMEs by focusing purely on machine-readable technical requirements.