deep-research

Orchestrate parallel web, codebase, and third-party code research with GoT-driven tasks.

1|Updated Jul 24, 2025
One-click install
npx skills add https://github.com/bumfo/spark-protobuf-backport --skill deep-research
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/bumfo/spark-protobuf-backport/tree/main/.claude/skills/deep-research
Command: npx skills add https://github.com/bumfo/spark-protobuf-backport --skill deep-research

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the fragmented, time-consuming, and often contradictory nature of manual technical research. It automates the orchestration of multiple specialized agents to conduct comprehensive investigations across web, local codebase, and third-party libraries, ensuring thoroughness and accuracy.

Core Features & Use Cases

  • Multi-Source Orchestration: Coordinates web, codebase, and third-party code analysis agents in parallel.
  • Graph of Thoughts Methodology: Applies structured thinking to decompose questions, score findings, and synthesize comprehensive answers.
  • Contradiction Resolution: Actively identifies and resolves conflicting information by cross-referencing primary sources.
  • Use Case: Investigate a subtle performance anomaly in a Spark application by comparing official documentation, the project's specific implementation, and the underlying library's source code, receiving a consolidated report with actionable insights.

Quick Start

To initiate a deep research task, simply ask a complex technical question, and the Skill will automatically decompose it and launch parallel agents. Example: Perform deep research on Spark's Catalyst optimizer, comparing its documented behavior with its actual implementation in Spark 3.2.1.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I automate technical research across multiple sources like documentation, code, and libraries?

Deep research orchestrates specialized agents in parallel to investigate web sources, local codebases, and third-party code simultaneously, synthesizing findings into a consolidated report with citations and resolved contradictions.

Can I compare official documentation against actual implementation in my codebase and dependencies?

Yes. Deep research uses Graph of Thoughts methodology to decompose complex questions, execute parallel investigations across web, codebase, and library sources, then cross-reference findings to identify and resolve conflicting information.

How do I investigate performance anomalies by examining framework behavior across multiple code layers?

Deep research coordinates web research, codebase exploration, and third-party code analysis in a single orchestrated task, comparing documented behavior with actual implementation to surface actionable insights about performance issues.

What's the best way to answer complex technical questions that require evidence from multiple sources?

Deep research decomposes multifaceted questions using Graph of Thoughts, launches parallel agents to gather evidence from documentation, local code, and libraries, then structures and scores findings to deliver rigorous, citation-backed answers.

Do I need to manually coordinate web searches, code exploration, and library lookups separately?

No. Deep research automates the orchestration of web browsing, codebase exploration, and code-lookup agents in parallel, eliminating fragmented manual research and ensuring comprehensive, contradiction-resolved investigation.

How does Graph of Thoughts improve the quality of technical research?

Graph of Thoughts structures investigation by decomposing questions into logical subtasks, scoring findings for relevance and accuracy, then synthesizing evidence across sources with ground-truth validation to produce coherent, well-supported conclusions.