Skill UI

A sophisticated system for recovering and reconstructing complex project context across distributed, multi-agent AI workflows. It utilizes advanced semantic vector search, relevance filtering, and optimized rehydration patterns to maintain high knowledge fidelity and ensure seamless continuity in long-running, complex AI projects. Essential for advanced RAG implementations.

Framework for building stateful, multi-actor AI applications with LangGraph, covering graph construction, reducers, persistence, human-in-the-loop patterns, and ReAct agents suitable for Python production systems.

Provides expert guidance for using llama-factory to fine-tune LLMs through its WebUI, covering QLoRA/LoRA quantized setups, multimodal models, and automation tips for developers exploring models like Llama, Qwen, or Gemma.

Analyze sexual health records, ED severity, STD screening, contraception effectiveness, and sexual activity trends. Useful for generating structured risk reports and linking sexual health data with medication, chronic disease, and lifestyle factors.

This technique is designed for entity recognition and hypothesis generation when identifying Indian or Bollywood actors based on indirect clues. Instead of running broad searches, users should hypothesize a small set of candidate names (1-3) using the provided clues (e.g., awards, family history, controversies). Subsequently, targeted verification searches must be executed for each constraint against the candidate names to confirm the correct identity.

This technique is designed to identify South Asian (Indian/Bollywood) films when the question provides insufficient information to name the movie directly. Instead, it requires chaining together the biographical details of the actor or director—such as birth year, debut era, family relations, or awards—to first uniquely identify the person, and then use that identity to locate the target film and verify all constraints.

This guide provides production-ready code patterns for integrating with the Cohere AI API v2 using TypeScript and Python. It covers essential best practices, including singleton client management, robust retry mechanisms, type-safe chat wrappers, streaming, embedding, reranking, and structured JSON output. Use this when establishing team coding standards or refactoring existing Cohere integrations.

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.

An advanced system that analyzes coding sessions via hooks to extract atomic, confidence-scored behaviors called "instincts." It tracks patterns, allowing users to evolve these instincts into comprehensive skills, commands, or agents. V2.1 introduces project-scoped learning, preventing knowledge contamination between different codebases and improving development workflow efficiency.

Provides comprehensive guidelines for designing, auditing, and documenting empirical evidence for high-level AI research submissions (e.g., AAAI). It covers essential methodologies including strong baselines, single-factor ablations, robustness testing (shift, seed, prompt), statistical significance, and ensuring perfect alignment with reproducibility checklists and ethical reporting. Use this guide before submission to ensure empirical evidence strongly supports all claimed AI contributions.

This guide details the complex, multi-stage review process for top-tier AI conferences like IJCAI. It covers strategies for Phase 1 summary rejection, managing conflicts of interest, understanding the roles of decision actors (PC, AC, SAC), and optimizing paper clarity to maximize acceptance chances. Use this when planning a submission strategy or interpreting reviewer feedback.

This tool serves as a pre-flight safety check for legal prompts before submission to third-party AI surfaces (e.g., ChatGPT, Claude). It assesses whether the content risks compromising attorney-client privilege or work-product protection. The output provides a structured safety band (SAFE/CAUTION/STOP), detailed triggered risk factors, a discovery impact summary, and a legally safe, redacted rewrite.