How long does a research task take?

The skill uses hard timeouts to ensure speed: Quick mode takes 2 minutes, Standard takes 3 minutes, and Extensive mode takes up to 10 minutes.

Where are the research findings stored?

Findings are first stored in a timestamped scratchpad directory for drafting and then moved to a permanent history archive with a README and executive summary.

What happens if one research agent is slow?

The skill follows strict timeout rules. Once the time limit is reached, the main agent stops waiting and synthesizes the final report using all available data gathered up to that point.

How many agents can run simultaneously?

Depending on the mode, you can launch 3 agents (Quick), 9 agents (Standard), or up to 24 agents (Extensive) to cover different angles of a research topic.

What is the Multi-Agent Research Assistant skill?

It is a specialized capability for Claude Code that allows it to launch multiple parallel agents (Perplexity, Claude, and Gemini) to perform deep-dive research and synthesize findings into a single report.

Multi-Agent Research Assistant

Name: Multi-Agent Research Assistant
Author: rafaelcalleja

byrafaelcalleja

•

网络抓取与数据收集

Orchestrates parallel research agents across Perplexity, Claude, and Gemini to deliver synthesized, multi-perspective reports with source attribution.

关于

The Research skill transforms Claude into a powerful information engine by deploying swarms of specialized agents to investigate topics simultaneously. By decomposing complex queries into sub-questions and running up to 24 parallel researchers, it provides comprehensive coverage that far exceeds standard search capabilities. The skill features three distinct modes—Quick, Standard, and Extensive—each with strict timeout rules to ensure rapid delivery of findings. Results are automatically structured, synthesized with confidence levels, and archived using a standardized history pattern for permanent reference.

主要功能

Strict timeout management to ensure prompt synthesis of partial results
Three research tiers: Quick (3 agents), Standard (9 agents), and Extensive (24 agents)
1 GitHub stars
Automated query decomposition for multi-domain coverage
Standardized workflow for scratchpad drafting and permanent history archiving
Multi-engine orchestration using Perplexity, Claude, and Gemini agents

使用场景

Comprehensive technical investigations for documentation and project planning
Rapid synthesis of current events from multiple academic and web sources
Deep-dive analysis of emerging technologies and market trends

关于

主要功能

Strict timeout management to ensure prompt synthesis of partial results
Three research tiers: Quick (3 agents), Standard (9 agents), and Extensive (24 agents)
1 GitHub stars
Automated query decomposition for multi-domain coverage
Standardized workflow for scratchpad drafting and permanent history archiving
Multi-engine orchestration using Perplexity, Claude, and Gemini agents

使用场景

Comprehensive technical investigations for documentation and project planning
Rapid synthesis of current events from multiple academic and web sources
Deep-dive analysis of emerging technologies and market trends