AI/ML

This skill provides a structured guide for implementing memory systems in AI agents, covering file storage, vector RAG, knowledge graphs, and temporal graphs. It includes clear activation triggers, architecture comparisons, and practical code examples for entity tracking and temporal queries.

Provides foundational knowledge for designing and debugging AI agent systems by explaining context components, attention mechanics, and progressive disclosure. Helps developers optimize token usage and manage system prompts, tool definitions, and message history effectively.

This skill provides patterns for diagnosing and fixing context degradation in LLM agents, focusing on issues like lost-in-middle effects and context poisoning. It offers concrete strategies like the four-bucket approach and architectural patterns to mitigate performance drops in long conversations.

A comprehensive pipeline for creating SFT datasets from books and training style-transfer models with detailed architectural guidance and practical implementation examples.

This skill wraps Google's Gemini CLI to provide AI-powered code analysis and generation within Claude. It offers multiple execution methods, configurable models via environment variables, and clear timeout handling. The Python script has zero dependencies and works cross-platform.

This skill implements an AI-driven email processing pipeline where LLMs handle all intent parsing and tool invocation, avoiding regex or state machines. It emphasizes conversation persistence, raw attachment passing to models, and learning from real bugs like 'no I mean...' misinterpretations.

A sophisticated skill for implementing BDI cognitive architectures, enabling RDF-to-mental-state transformations with strong ontological foundations and neuro-symbolic integration patterns.

Provides structured patterns for designing multi-agent LLM systems to overcome single-agent context limits. Covers supervisor, peer-to-peer, and hierarchical architectures with concrete implementation examples. Includes memory layers and coordination protocols for distributing complex tasks across specialized agents.

This skill implements an adaptive learning system for AI agents using ReasoningBank. It enables agents to record experiences, recognize patterns, optimize strategies, and continuously improve through meta-learning. The tool integrates with AgentDB for persistence and supports vector search for pattern matching.

This skill provides detailed guidance for using scikit-learn, covering classification, regression, clustering, and preprocessing. It includes practical code examples, algorithm selection tables, and warnings about common mistakes like data leakage. The documentation helps users build ML pipelines and evaluate models effectively.

This skill enables coordination of multiple AI agents for parallel task execution using agentic-flow. It supports mesh, hierarchical, and adaptive topologies with features like load balancing, fault tolerance, and shared memory. Useful for complex projects requiring distributed AI workforces.

Provides detailed instructions for using OpenAI Whisper to transcribe audio/video files into multiple formats including SRT, VTT, and JSON. Covers model selection, audio extraction, timing synchronization, and includes batch processing scripts. Focuses on practical implementation with performance optimization tips.