What is the difference between Quick Search and Deep Research?

Quick Search is intended for fast factual lookups and definitions (2-5 seconds), while Deep Research provides a comprehensive analysis of complex topics that can take several minutes to complete.

Does this skill require a Parallel AI API key?

Yes, you must obtain an API key from the Parallel AI platform (platform.parallel.ai) to enable the search and task capabilities.

Is permission required for every search?

Quick searches are performed automatically when needed, but the skill is configured to always ask for user permission before initiating a Deep Research task due to higher resource usage.

Will the agent wait for deep research to finish?

No, the skill implements a non-blocking design using the MatClaw scheduler. The agent starts the task and polls for results in the background, allowing it to remain responsive.

Parallel AI Research for MatClaw

Name: Parallel AI Research for MatClaw
Author: DingyangLyu

byDingyangLyu

•

Data Science & ML

Integrates Parallel AI MCP servers into MatClaw to enable high-speed web lookups and comprehensive deep research capabilities.

This skill enhances the MatClaw materials-science agent by integrating advanced web research capabilities via Parallel AI's MCP servers. It provides two distinct research modes: a fast 'Quick Search' for factual lookups and a 'Deep Research' mode for exhaustive analysis. The integration uses a non-blocking scheduler pattern, ensuring that long-running research tasks don't hang the agent's execution environment. By automating the configuration of environment variables, MCP servers, and tool permissions, this skill allows MatClaw to effectively bridge the gap between simulation workflows and the latest scientific literature available online.

Key Features

01Quick factual lookups via Parallel Search API

02Built-in permission gating for high-resource research tasks

03Automated MCP server and environment configuration

04Comprehensive deep research through Parallel Task API

05Non-blocking polling architecture using MatClaw scheduler

0632 GitHub stars

Use Cases

01Verifying current news, events, or citations for research workflows

02Conducting in-depth historical research on specific scientific concepts

03Retrieving the latest material properties or experimental data from the web

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dingyanglyu/matclaw add-parallel

For use in Claude.ai and ChatGPT