Skill UI

A robust protocol designed for multi-agent collaboration on long-running projects. It simulates a self-optimizing city where diverse AI agents (citizens) interact using decentralized markdown files (The Register, Contracts, Chronicle). This framework enables cross-vendor optimization, dynamic work routing, and constitutional evolution without requiring a centralized server or specific runtime environment.

A comprehensive guide for training or fine-tuning advanced sentence-transformer models. This skill assists in selecting the correct architecture—including Bi-encoders (SentenceTransformer), CrossEncoders (for reranking), and SparseEncoders—for various Natural Language Processing (NLP) tasks. Use cases include semantic retrieval, similarity matching, document clustering, and pair-wise classification.

This skill provides expert capabilities for training and fine-tuning transformer language models using the TRL (Transformers Reinforcement Learning) library. It supports state-of-the-art post-training methods, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and Reward Model training. Use this skill to align and customize foundation models for advanced tasks via CLI commands.

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.

A sophisticated, dark-themed presentation template designed for AI-native projects, knowledge graphs, or deep technical deep dives. It features a gradient cover with floating orbs and an interactive SVG force-directed graph, complemented by JetBrains Mono code highlighting and modern glassmorphism elements for section slides. Perfect for tech talks and product launches.

A technical specialization focused on Generative Engine Optimization (GEO). This skill ensures that specific web pages are structured, documented, and optimized for modern AI answer engines (Google AI Overviews, ChatGPT, Bing Copilot). It focuses on technical elements like structured data (JSON-LD), semantic HTML, and robust metadata to maximize discoverability, trustworthiness, and quotability for AI consumption, rather than traditional SEO ranking.

This guide details how to effectively delegate complex, multi-step tasks to various specialized AI agents (Codex, Pi, Claude, Hermes). It covers best practices for maintaining conversation context, handling TUI prompts (e.g., avoiding newlines, using quotes), techniques for reliable polling, and choosing the optimal agent for coding, design, or general purpose work. Essential for advanced AI workflow automation.

The /goal command converts standard agent prompts into persistent, self-correcting loops (plan → act → test → review → iterate). It is designed for long-running, complex engineering tasks like migrations, coverage lifts, or refactoring, ensuring progress is tracked against a verifiable stop condition and mandated documentation.

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 comprehensive guide outlines best practices for building reliable and robust AI agent harnesses. It covers optimizing action spaces, defining structured tools, managing observation formats, and implementing essential error recovery contracts. Use these principles to dramatically improve agent completion rates, manage context budgets effectively, and transition AI prototypes into stable, production-grade autonomous systems.

This skill provides a systematic, four-phase workflow for diagnosing and recovering from AI agent failures. It guides the agent to first capture the failure state, diagnose the root cause (e.g., loop, context overflow, state drift), apply minimal containment recovery actions, and finally generate a detailed introspection report, minimizing the need for human intervention.