Skill UI

SAELens provides a framework for training and analyzing Sparse Autoencoders (SAEs). SAEs decompose the dense, often polysemantic activations of large language models into sparse, monosemantic features. Use this when you need to discover the discrete, interpretable concepts a model has learned, study feature superposition, or analyze specific safety-relevant behaviors (like bias or deception) within deep neural networks.

Vexor is a powerful command-line interface designed for advanced semantic search across local files. By utilizing vector embeddings and integrating with large language models (like Claude/Codex), it allows users to search files based on meaning and context, rather than just exact keywords. This is ideal for large codebases or document repositories where traditional keyword search methods are insufficient.

This comprehensive skill provides an end-to-end solution for video and audio content creation. Users can upload videos from various sources (YouTube, local files) and perform advanced operations such as real-time streaming, semantic search, and complex editing (trimming, overlays). Key features include AI-powered transcription, subtitle generation, and the ability to generate diverse media assets like images, music, sound effects, and voiceovers, making it ideal for content creators and media professionals.

A powerful tool for conducting comprehensive academic research and literature reviews. It searches, analyzes, and synthesizes key ideas from multiple sources, including local PDF libraries (Zotero, Obsidian), the general web, and specialized academic databases (Semantic Scholar, DeepXiv). Use this when you need to map the historical context, identify related works, or understand complex concepts across multiple studies on a given topic.

OpenViking Memory enables AI agents to recall and persist long-term semantic knowledge using MCP tools. It supports retrieving prior context at task start via find/search/read and storing durable facts, preferences, and lessons via remember. Ideal for coding, debugging, and multi-step workflows, ensuring continuity across sessions without automatic lifecycle hooks.

Instrument LLM applications with OpenInference semantic conventions in Phoenix. Covers setup, auto and manual instrumentation, 9 span types, session tracking, production batching and PII masking for Python and TypeScript.

Provides comprehensive intelligence access across news, finance, and media. Users can analyze news articles for media bias and balanced perspectives across thousands of sources. It also delivers live data for stocks, ETFs, and crypto, offering AI-generated bull/bear cases, advanced options pricing, and semantic meme searching capabilities.

Master the complexities of building robust Retrieval Augmented Generation (RAG) pipelines using LangChain. This guide addresses common pitfalls in vector search, such as score semantics differences (L2 vs. Cosine), embedding dimension mismatches, and chunking issues for code/Markdown. Learn best practices for selecting between FAISS, Pinecone, Chroma, and PGVector, normalizing scores, and implementing hybrid keyword+vector search for maximum retrieval quality.

Implement robust human-in-the-loop (HITL) approval gates for sophisticated LangGraph workflows. This skill enables the graph to pause execution before critical, irreversible actions (like sending emails or calling tools). It manages complex state serialization, provides clear resume semantics using `Command(resume=...)`, and integrates with external UIs (e.g., Slack) to await human decision, ensuring safe and controlled workflow execution.

Qdrant is a high-performance vector database. This client SDK suite provides official, easy-to-use libraries for multiple programming languages (including Python, Java, Go, and TypeScript). It enables developers to seamlessly integrate sophisticated vector search capabilities into their applications, supporting both REST and gRPC APIs. Ideal for building semantic search engines, recommendation systems, and advanced AI applications.

A self-hosted, semantic memory service designed for AI agents. It allows agents to save worklogs, decisions, and notes, and recall this knowledge across sessions based on meaning rather than exact keywords. Using pgvector and structured workspaces, it maintains persistent, context-aware knowledge bases, making it ideal for collaborative AI projects.

This skill provides a comprehensive guide for building scalable Retrieval-Augmented Generation (RAG) pipelines and persistent agent memory using Pinecone as the vector database. It covers index setup, embedding large corpuses, implementing dense and hybrid semantic search, and advanced metadata filtering for real-world production systems. Use this for building searchable knowledge bases and context-aware AI agents.