Skill UI

A powerful plugin enabling Claude Code to delegate complex tasks to external AI CLIs (such as OpenAI Codex or Google Antigravity/Gemini) without exiting the conversation. It supports multi-model workflows, allowing users to seek a second opinion, cross-check deep analysis, or delegate side tasks (like research or code review). The core value is that Claude synthesizes the raw outputs, providing a critical review and reconciliation of findings across all consulted models.

This skill integrates with DeepAPI to facilitate advanced web scraping, market research, and controlled email workflows. It handles scraping data from various sources like websites, LinkedIn, GitHub, and Twitter, while managing costs and ensuring explicit user approval for sensitive actions like sending emails. Use it for robust data extraction and automated reporting.

A comprehensive suite designed for AI agents to perform safe, deep, and multi-format web searches. It allows agents to retrieve current web information, analyze PDFs, extract YouTube transcripts, and access GitHub code files. Instructions include guidelines for varying search depths and using specialized tools for code context retrieval.

This skill converts ambiguous or general research needs into a single, highly precise, self-contained prompt. It guides the user to define the project context, the core question, and specific sub-questions, ensuring the resulting prompt is rigorous enough for deep, multi-source AI research, minimizing ambiguity and guesswork.

This skill executes the DeepSWE benchmark, a comprehensive coding-agent evaluation suite. It uses OpenRouter and mini-swe-agent to score various Large Language Models (LLMs) on 113 tasks, providing reproducible results for model comparison, development assessment, and research benchmarking. Use this when a standardized, deep evaluation of LLM coding capabilities is required.

This skill provides a unified interface to query and manage complex knowledge stored in AgentDB. It facilitates advanced retrieval mechanisms, including semantic routing, hierarchical recall, causal graph construction, and context synthesis. Use it when you need to reliably store, retrieve, or search interconnected knowledge across multiple agent sessions, ensuring high fidelity and deep contextual understanding.

This advanced workflow is designed to generate comprehensive, fact-checked research reports on any complex topic. It systematically performs fan-out web searches across multiple defined angles, fetches multiple sources, and employs a rigorous adversarial verification system (3-vote) to validate claims. The final output is a highly cited, synthesized report, ideal for academic or critical business research.

A mechanism for spawning hierarchical, nested sub-agents up to five levels deep, designed specifically for deep task delegation while managing context limitations. Instead of parallel execution, this ensures that each layer of decomposition receives a fresh context window, preventing the parent agent's context budget from being overwhelmed by complex, multi-step workflows. Ideal for research, verification, and synthesis processes.

This is an advanced, structured workflow designed to conduct deep, multi-source investigations. It synthesizes current public evidence with user-provided context to generate highly structured reports. The workflow explicitly differentiates between verifiable facts, user-supplied evidence, logical inferences, and actionable recommendations, making it ideal for complex comparative analysis or continuous monitoring tasks.

This tool synthesizes deep, accumulated research findings stored in memory into structured, actionable reports. It systematically grades evidence quality (High, Medium, Low), identifies and resolves conflicting claims, and generates comprehensive outputs including an executive summary, ranked key findings, limitations, and concrete recommendations. Ideal for transforming raw research data into strategic insights.

Dispatches computationally intensive quantitative financial jobs (such as multi-year walk-forward backtesting, large Monte-Carlo simulations, parameter sweeps, and deep model training using neural-trader) to a managed cloud agent. This is necessary for workloads that exceed local computational limits, providing robust, scalable infrastructure for professional quantitative research.