10x-research

Research codebases by spawning parallel sub-agents and synthesizing findings into structured documents.

Updated Aug 26, 2026
One-click install
npx skills add https://github.com/michaail/hybrid-logs-analyzer --skill 10x-research-michaail
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: 10x-research
Source: https://github.com/michaail/hybrid-logs-analyzer/tree/main/.cursor/skills/10x-research
Command: npx skills add https://github.com/michaail/hybrid-logs-analyzer --skill 10x-research-michaail

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Answering broad questions about a large codebase requires reading many files and connecting patterns across components, which is slow and error-prone when done sequentially. This Skill orchestrates parallel sub-agents to investigate different research areas concurrently and synthesizes their findings into a single, well-structured research document. ## Core Features & Use Cases - Parallel sub-agent research: Spawns 2-4 concurrent Explore and general-purpose agents, each focused on a specific research dimension with file:line references. - Scope clarification: Asks targeted questions about scope, depth, and focus before launching agents when the query is ambiguous. - Structured research documents: Generates a Markdown research artifact with YAML frontmatter, code references, architecture insights, and historical context under context/changes/<change-id>/research.md. - Use Case: Ask "how does authentication work across the API and worker services" and receive a synthesized document with concrete file:line evidence, architectural patterns, and links to prior decisions. ## Quick Start Invoke the 10x-research skill and ask it to research how a specific feature or system works across the codebase.

Frequently Asked Questions about 10x-research

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

FAQPage Schema
How do I research a large codebase with AI agents?

Invoke this skill with a research question, and it decomposes the query into research areas, spawns 2-4 parallel sub-agents to investigate concurrently, then synthesizes their findings into a structured document with file:line references.

How to answer architecture questions about unfamiliar code?

Provide a focused question about the system or feature, and the skill coordinates Explore agents for fast file and pattern searches plus general-purpose agents for deep multi-file analysis, then connects findings across components.

What output does the codebase research produce?

It writes a Markdown research document at context/changes/<change-id>/research.md containing YAML frontmatter, a summary, detailed findings, code references, architecture insights, historical context, and open questions.

Can I ask follow-up questions after the research completes?

Yes. Follow-up questions append new sections to the same research document, update the last_updated frontmatter fields, and may spawn additional sub-agents for further investigation.

When should I not use parallel sub-agent research?

Avoid it for trivial lookups answerable by reading one or two files directly. The multi-agent workflow is designed for broad, cross-component questions where decomposition and concurrent exploration provide real value.