deep-research

Conduct multi-source web research and synthesize findings into cited reports with source credibility scoring.

Updated Jul 10, 2026
One-click install
npx skills add https://github.com/SaberAloui/agri-hub --skill deep-research-saberaloui
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/SaberAloui/agri-hub/tree/main/.local/secondary_skills/deep-research
Command: npx skills add https://github.com/SaberAloui/agri-hub --skill deep-research-saberaloui

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Answering complex questions that require more than a quick search—market analysis, technology evaluations, due diligence, fact-checking—demands gathering many sources, cross-checking conflicting claims, and producing a well-cited report, which is slow and error-prone when done manually. ## Core Features & Use Cases - Parallel Multi-Agent Research: Decomposes a topic into focus areas and launches simultaneous research subagents, each running multiple web searches and fetching full articles. - Source Triangulation & Credibility Tiers: Scores sources as Tier 1 (government/academic), Tier 2 (major publications), or Tier 3 (blogs/promotional), requiring 3+ sources per major claim and resolving conflicting data transparently. - Structured Report Generation: Produces prose-first reports with executive summary, key findings, analysis, limitations, recommendations, and a numbered source list with inline citations. - Use Case: Ask for a deep dive on the state of the electric vehicle market in 2026, and receive a comprehensive report covering market share, battery technology, policy, infrastructure, and consumer economics with 25+ cited sources. ## Quick Start Research the competitive landscape of the electric vehicle market in 2026 and write a cited report with an executive summary and recommendations.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I run a deep research report on a complex topic?▼

State your research question and desired depth, such as a standard or deep dive. The skill decomposes the topic into focus areas, launches parallel research subagents to gather sources, triangulates the findings, and writes a cited report with an executive summary and recommendations.

What research depth tiers are available and when should I use each?▼

Three tiers exist: Quick (3 subagents, 8+ sources, 2-5 minutes) for focused questions, Standard (5 subagents, 15+ sources) for most research requests, and Deep (5 plus 2 gap-fill subagents, 25+ sources) for critical decisions and comprehensive reviews.

How does the research verify claims and handle conflicting sources?▼

Every major claim requires support from at least three sources, with credibility tiers ranking government and academic sources above publications and blogs. When sources conflict, higher-tier and more recent data is preferred, and discrepancies are explicitly noted rather than silently resolved.

When should I use a simple web search instead of deep research?▼

Use a direct web search for simple factual lookups answerable in one or two queries. Deep research is intended for multi-angle analysis, and it defers stock ticker analysis to stock-analyzer, codebase searches to grep, and user-provided datasets to data-visualization.

What are the limitations of web-based deep research?▼

It cannot access paywalled journals, subscription databases, or social media platforms, and webFetch may fail on JavaScript-heavy or paywalled pages. It also cannot conduct primary research such as surveys or interviews, and findings are a snapshot in time.