skills.rest
korchasa
korchasaCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

flow-skill-deep-research

Plan and execute multi-directional web research with sequential sub-agents.

3|Updated Oct 5, 2025
Search Tags:#research#reporting#synthesis#citation#web-search#mult-agent
One-click install
npx skills add https://github.com/korchasa/flow --skill flow-skill-deep-research
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: flow-skill-deep-research
Source: https://github.com/korchasa/flow/tree/main/framework/skills/flow-skill-deep-research
Command: npx skills add https://github.com/korchasa/flow --skill flow-skill-deep-research

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enables rigorous, citation-backed web research by orchestrating multi-directional searches with sequential sub-agents and synthesizing findings into a structured report.

Core Features & Use Cases

  • Planner decomposes a topic into non-overlapping research directions and defines success criteria.
  • Worker sub-agents perform direction-specific searches, evaluate sources for authority, fetch content, and save outputs.
  • Synthesizer reads all findings, merges results, and produces an executive summary plus a full markdown report with citations.
  • Built-in search tool support (e.g., web search), Playwright-based fetch options, and MCP-style search tools.
  • Output is a traceable, source-backed report suitable for decisions, policy, or scholarly work.

Quick Start

Provide a topic and it will plan directions, run staged web searches, synthesize findings with citations, and save a final report.

Frequently Asked Questions about flow-skill-deep-research

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

FAQPage Schema
How do I automate multi-directional web research with inline citations?▼

Multi-directional web research is automated by decomposing topics into directions, running sequential sub-agents for structured searches, and synthesizing findings into a markdown report with inline citations.

How does an agent evaluate source authority during web searches?▼

Source authority is evaluated during web searches by worker sub-agents assessing fetched content against defined success criteria, filtering out low-quality sources before synthesis.

Can I use MCP-style search tools for citation-backed reporting?▼

MCP-style search tools can be used for citation-backed reporting, alongside built-in search and Playwright-based fetch options, to gather and verify source content for the final markdown report.

What is the best way to synthesize findings from multiple search directions?▼

The best way to synthesize findings from multiple search directions is using a synthesizer agent to merge results, producing an executive summary and a full markdown report with traceable citations.

Do I need web search and verification tools to run multi-agent research?▼

You need web search and verification tools to run multi-agent research, as the sequential sub-agents depend on them to fetch content, evaluate sources, and compile traceable findings.

Related Skills

Explore More Education & Research Skills→
lvndrylvndry
47

deep-research

Automate multi-source research with decomposition, verification, and synthesis.

Community
Advanced
joneqianjoneqian

deep-research-pro

Automate multi-source web research and generate cited reports with executive summaries.

Community
Advanced
Zenobia000Zenobia000
20

sunnydata-deep-research

Automate multi-source web research and synthesize cited reports.

Community
Advanced
ClementATHClementATH

external-context

Decompose user queries into parallel search facets and coordinate document-specialist agents to fetch cited external references.

Community
Advanced