web-research

Organize multi-source web research into local cited files via subagents.

Updated Mar 8, 2026
One-click install
npx skills add https://github.com/D26FORWARD/MatrixFin --skill web-research-d26forward
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: web-research
Source: https://github.com/D26FORWARD/MatrixFin/tree/main/deepagents/libs/cli/examples/skills/web-research
Command: npx skills add https://github.com/D26FORWARD/MatrixFin --skill web-research-d26forward

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Organizes and streamlines multi-source web research by enforcing planning, delegating subtopics to controlled subagents, and consolidating findings into local, cited files.

Core Features & Use Cases

  • Plan-first workflow that requires a research folder and a written research_plan.md outlining 2–5 subtopics.
  • Spawn task-based research subagents that run up to 3–5 web searches each and write findings to local files for later synthesis.
  • Use Case: Conduct a comparative analysis of competing technologies by delegating distinct comparison dimensions to parallel subagents and synthesizing their findings into a final report with source URLs.

Quick Start

Create a research folder for your topic, write a research_plan.md with clear subtopics, and spawn task subagents to save findings into the folder for synthesis.

Frequently Asked Questions about web-research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I conduct structured web research with multiple sources?

Structured web research uses a plan-first workflow to delegate 2-5 subtopics to parallel subagents, saving findings to local files for deterministic synthesis and citation.

How do I organize comparative analysis findings into local research files?

Organize comparative analysis findings by creating a research folder, writing a research_plan.md, and spawning task subagents that save their findings locally for final synthesis.

Can I use parallel subagents for multi-source fact-finding?

Yes, you can spawn task-based research subagents for multi-source fact-finding, with each subagent limited to 3-5 web searches before writing results to local files.

What's the best way to synthesize current events from web search results?

Synthesize current events by enforcing a plan-first workflow that delegates research to controlled subagents, consolidating their local file outputs into a cited final report.

How many web searches can each research subagent perform?

Each task-based research subagent is limited to 3-5 web searches to ensure controlled information gathering before writing findings to local files for synthesis.

Do I need a research plan before starting web research?

Yes, a written research_plan.md outlining 2-5 subtopics is required before spawning subagents, ensuring structured delegation and deterministic synthesis of findings.