Skill UI

This toolset allows users to save and restore complete agent sessions, ensuring context continuity across multiple conversations. It captures the agent's entire state, including memory, configuration, task progress, and learning metrics. Use it when pausing work and needing to resume later with full context, or for checkpointing long-running tasks.

This skill executes comprehensive historical backtesting using the high-performance neural-trader engine. It supports features like walk-forward validation and captures key financial metrics (Sharpe, Drawdown). Critically, it signs the resulting artifact using Ed25519 cryptography, providing non-repudiable tamper evidence required for promoting research models to live trading systems.

This powerful tool utilizes Playwright to record high-quality, professional web application demo videos. It simulates natural user interactions, injecting a visible cursor overlay and adhering to specific pacing guidelines. It is ideal for creating comprehensive tutorials, feature walk-throughs, or documentation for stakeholders.

A comprehensive skill for ingesting, analyzing, and manipulating multi-modal video and audio content. Supports real-time desktop capture, live stream monitoring (RTSP), and file uploads (URL/local). Capabilities include advanced indexing (visual, semantic, temporal), precise moment searching, professional timeline editing (subtitles, overlays, dubbing), transcoding, and generating actionable alerts. Ideal for content creators, monitoring systems, and complex media workflows.

This helper tool is designed to assist with a wide variety of tasks and needs. It encourages users to use their own judgment and judgment to handle diverse prompts, making it a flexible and adaptable resource for general problem-solving scenarios. Ideal for brainstorming, general queries, and initial exploration.

This skill acts as an advanced orchestration protocol for reviewing GitHub pull requests. It is stack-agnostic, meaning it does not enforce its own rules. Instead, it dynamically discovers the project's existing standards—such as the test runner, architectural conventions, and requirement specifications—before running six specialized subagents in parallel. It provides a comprehensive, context-aware review and summary directly on GitHub.