deep-research

Investigate complex topics and produce evidence-backed research reports with coverage and uncertainty.

4|Updated Feb 23, 2026
One-click install
npx skills add https://github.com/npow/claude-skills --skill deep-research-npow
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/npow/claude-skills/tree/main/deep-research
Command: npx skills add https://github.com/npow/claude-skills --skill deep-research-npow

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill replaces ad hoc, shallow research with a systematic workflow that explores a topic across independent dimensions, tracks coverage, evaluates evidence quality, and clearly reports gaps and uncertainty.

Core Features & Use Cases

  • Multi-Dimensional Exploration: Investigate applicable WHO, WHAT, HOW, WHERE, WHEN, WHY, and LIMITS dimensions alongside cross-cutting concerns such as prior failures, baselines, actual usage, and strategic timing.
  • Parallel Evidence Gathering: Coordinate research agents across prioritized directions with controlled depth, deduplication, watchdogs, structured findings, and bounded budgets.
  • Quality-Controlled Synthesis: Produce sourced reports with claim registers, corroboration levels, contradiction analysis, citation spot-checks, stale-source warnings, coverage assessments, and known gaps.
  • Use Case: For a question about how vector databases handle updates, the Skill can map the technical mechanisms, production constraints, competing approaches, maturity, failure modes, and evidence supporting each conclusion.

Quick Start

Use the deep-research skill to investigate the topic of how vector databases handle updates and produce a comprehensive, source-backed report.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I conduct systematic investigation for technical research topics?

Systematic investigation maps technical topics across independent dimensions like mechanisms and constraints, coordinates parallel evidence gathering, and produces structured reports with deduplication and source-quality assessment. It replaces ad hoc research with bounded multi-agent execution tracking coverage and uncertainty.

What is the best way to handle source verification and contradiction analysis in evidence synthesis?

Source verification in evidence synthesis applies citation spot-checks, evaluates evidence quality, and explicitly resolves contradictions across findings. The process generates a claim register with corroboration levels to ensure conclusions are backed by validated sources rather than unsupported assertions.

How does parallel evidence gathering work for comprehensive exploration?

Parallel evidence gathering coordinates multiple research agents across prioritized directions with controlled depth. It enforces bounded budgets, uses watchdogs for deduplication, and structures findings to ensure multi-dimensional exploration covers WHO, WHAT, HOW, WHERE, WHEN, WHY, and LIMITS without redundancy.

Can I use this approach for comparative analysis of competing technical approaches?

Yes, comparative analysis of competing approaches fits this systematic investigation framework. It maps technical mechanisms, production constraints, maturity levels, and failure modes side-by-side, applying coverage analysis to ensure all relevant cross-cutting concerns and prior failures are evaluated with sourced evidence.

How do I report unexplored areas and known gaps in a deep research report?

Reporting unexplored areas requires explicit coverage assessments and honest documentation of known gaps within the research output. The synthesis process tracks stale-source warnings and unexplored dimensions, ensuring the final report transparently communicates uncertainty and boundaries of the investigation.

When do I need coverage analysis for open-ended research?

Coverage analysis is needed for open-ended research when exploring complex topics across multiple dimensions to prevent shallow conclusions. It tracks whether applicable WHO, WHAT, HOW, WHERE, WHEN, WHY, and LIMITS factors are sufficiently investigated, ensuring comprehensive exploration before final evidence synthesis.