Skill UI

This guide provides detailed best practices for ensuring the reproducibility of advanced AI research across various domains, including algorithms, theories, multi-agent systems, and datasets. It outlines the specific evidence required—from pseudocode and seeds to full proofs and public datasets—to elevate a submission's credibility rating from 'irreproducible' to 'convincing' for top-tier AI conferences.

A comprehensive guide for authors targeting the International Conference on Autonomous Agents and Multiagent Systems (AAMAS). This skill helps authors strategically position their work by focusing on core themes like autonomous agents, multiagent learning, game-theoretic interaction, and decentralized decision-making. It covers submission framing, evidence requirements, and risk assessment to ensure manuscripts align with the venue's technical focus.

This skill allows agents to retrieve and inject relevant operational memories from the OpenViking experience library into the current prompt context. It is designed for complex, multi-step tasks involving tool usage, code operations, or decision-making workflows. By searching and reading past successful executions, the agent improves context awareness and decision accuracy, moving beyond simple Q&A.