SciVisAgentSkills: Design and Evaluation of Agent Skills for Scientific Data Analysis and Visualization
For researchers and practitioners in scientific visualization, this work provides a structured approach to augment coding agents with domain-specific expertise, enabling more reliable long-horizon workflows.
The paper introduces SciVisAgentSkills, a collection of reusable agent skills that enhance coding agents for scientific visualization tasks, and evaluates them on a benchmark of 108 multi-step tasks. Results show improved mean task scores and token efficiency, though benefits depend on the agent harness and tool setting.
Recent advances in agentic visualization have enabled the translation of natural language into executable scientific visualization (SciVis) workflows. While general-purpose coding agents show strong capabilities, they often lack the tool-specific expertise required for SciVis tasks. In this work, we present SciVisAgentSkills, a collection of reusable agent skills that augment coding agents for scientific data analysis and visualization by encoding environment assumptions, tool usage patterns, and domain heuristics across scientific tools such as ParaView, napari, VMD, and TTK. We evaluate these skills on Codex and Claude Code using SciVisAgentBench, a benchmark of 108 expert-designed multi-step tasks. Results show that agent skills improve mean task scores across the evaluated suites, with token-efficiency benefits that depend on the agent harness and tool setting. These findings highlight the importance of structured procedural knowledge for enabling reliable, long-horizon SciVis workflows, while also showing that skills should be studied alongside the execution harness that loads and applies them. The skills are available at https://github.com/KuangshiAi/SciVisAgentSkills.