search-synth

Coordinate multi-agent web research with cross-fact-checking into a verified synthesis.

9|Updated Apr 2, 2026

Search Tags:#fact-checking #research #verification #multi-agent #synthesis #web

Ready to run

npx skills add https://github.com/dazzaji/interlateral_agents --skill search-synth

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: search-synth
Source: https://github.com/dazzaji/interlateral_agents/tree/main/.claude/skills/search-synth
Command: npx skills add https://github.com/dazzaji/interlateral_agents --skill search-synth

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Conducts thorough multi-agent web research with structured cross-fact-checking to deliver a coherent, verified synthesis.

Core Features & Use Cases

Orchestrates a MANAGER-led search with multiple SEARCHER agents to gather diverse sources.
Performs phase-driven workflows (Search, Fact-Check, Best and Verified Synthesis) to produce a reliable final report.
Produces an actionable, publishable synthesis suitable for decision-making and knowledge sharing.

Quick Start

In your environment, invoke search-synth on a topic and specify the participating agents and output file to generate a complete Best and Verified report.

Frequently Asked Questions about search-synth

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

FAQPage Schema

How do I conduct multi-agent web research with fact-checking and synthesis?▼

To conduct multi-agent web research with fact-checking, a MANAGER orchestrates multiple SEARCHER agents with web-search capability to gather sources, which a CHIEF_AUTHOR then synthesizes into a verified report using a three-phase workflow.

What is a verified synthesis report and how does cross-fact-checking work?▼

A verified synthesis is a coherent report created by cross-fact-checking facts across multiple SEARCHER agents before a CHIEF_AUTHOR crafts the final output. This ensures the resulting synthesis is reliable for decision-making and knowledge sharing.

How do I set up agents for multi-agent research synthesis?▼

Setting up multi-agent research synthesis requires configuring a MANAGER agent to lead, multiple SEARCHER agents with web-search capability, and a CHIEF_AUTHOR agent. You specify these participating agents and an output file when invoking the process on a topic.

Can I generate a publishable synthesis report from web research for decision-making?▼

Yes, you can generate a publishable synthesis report suitable for decision-making by running the phase-driven workflow. It produces an actionable, verified synthesis output written to a specified OUTPUT_FILE for knowledge sharing.

What are the limitations of using a multi-agent approach for research verification?▼

A limitation of this multi-agent research verification approach is its strict dependency on having a MANAGER, multiple SEARCHER agents with web-search capability, and a CHIEF_AUTHOR configured. It applies only when rigorous cross-fact-checking across multiple agents is required.