Skill UI

An expert system for designing and implementing advanced AI context architectures. It specializes in integrating intelligent memory systems, vector databases, and knowledge graphs to ensure AI applications maintain coherent, relevant, and scalable state across complex, multi-agent workflows. Essential for enterprise RAG implementation and stateful AI development.

This skill provides methodologies for designing sophisticated, layered memory architectures for AI agents. It moves beyond simple context windows and basic vector stores by incorporating short-term, long-term, and graph-based memory systems. Implementing this is crucial for agents that require persistence across sessions, maintaining complex entity consistency, and performing advanced, relationship-driven reasoning over accumulated knowledge bases, including time-aware fact tracking.

A persistent, graph-based knowledge base designed to accumulate structured knowledge throughout the entire research lifecycle. It stores and connects diverse entities—including published papers, novel ideas, conducted experiments, and testable claims. Inspired by LLM Wiki patterns, it ensures knowledge compounding, meaning the wiki becomes progressively smarter with every piece of input, rather than requiring re-derivation for every query.

This technique is used when the required answer in a case query is not directly stated but must be deduced by following a sequence of connected roles, people, or events. It involves mapping complex relationships, such as Victim -> Defendant -> Lawyer -> Witness, to trace indirect connections and reach the final piece of information. Ideal for advanced investigative research and deep knowledge graph traversal.

TorchDrug is a modular PyTorch library for graph learning tasks including molecular property prediction, molecule generation, and knowledge graph reasoning. It provides datasets, models, and training engines for building workflows in computational chemistry and biology. Ensure compatibility with PyTorch 1.8-2.0 and Python 3.7-3.10 before implementation.

This skill builds a persistent knowledge base combining LLM wiki patterns with structured graphs. It ingests documents once to generate wiki pages and extract entities/relations, avoiding re-derivation. It supports agents with evidence-based querying, wiki search, and knowledge accumulation, ensuring cross-references are never lost.

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.

Query and manage AgentDB knowledge base using controller bridge tools. Supports semantic routing, hierarchical recall, causal graphs, and context synthesis for agent sessions. Enables storing, retrieving, and searching structured knowledge across tasks.

A state-of-the-art memory search engine providing multi-faceted retrieval strategies. It supports default dense semantic matching, advanced hybrid search (sparse + dense with RRF), deep multi-hop Graph RAG for complex reasoning, and 'Smart' search utilizing query expansion, recency boosting, and MMR diversity reranking. Ideal for retrieving nuanced, diverse, and time-sensitive information from large, complex knowledge bases across multiple namespaces.

This comprehensive guide outlines best practices for achieving rigorous reproducibility in complex AI and knowledge graph research. It details how to pin every component—from tokenizers and KGs to LLM prompts and environment settings—ensuring that results are fully documented and verifiable by peers, even when proprietary data is involved.

This skill provides a comprehensive guide for authors targeting the International Semantic Web Conference (ISWC). It details the best practices for framing semantic web, knowledge graph, and ontology papers. It helps authors ensure their manuscript meets the venue's high standards regarding semantic interoperability, real-world data application, and specific contribution types, optimizing the submission strategy.

Compiles documents, notes, web pages, transcripts, research materials, or code repositories into an evidence-grounded, visualization-ready knowledge graph of typed entity nodes with statement-level provenance and typed, directional relationship edges.