One-click install
npx skills add https://github.com/ed3dai/ed3d-plugins --skill doing-a-simple-two-stage-fanout
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: doing-a-simple-two-stage-fanout
Source: https://github.com/ed3dai/ed3d-plugins/tree/main/plugins/ed3d-basic-agents/skills/doing-a-simple-two-stage-fanout
Command: npx skills add https://github.com/ed3dai/ed3d-plugins --skill doing-a-simple-two-stage-fanout

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill tackles the challenge of analyzing massive text, code, or data corpuses that exceed the context window of a single AI agent, ensuring thoroughness and accuracy.

Core Features & Use Cases

  • Distributed Analysis: Divides large datasets into manageable segments for parallel processing by multiple 'Worker' agents.
  • Quality Assurance: Employs 'Critic' agents to review Worker outputs, identifying gaps, inconsistencies, and cross-segment patterns.
  • Synthesis: A 'Summarizer' agent consolidates all reviews into a comprehensive final report.
  • Use Case: Analyzing a multi-gigabyte codebase for security vulnerabilities, where each Worker analyzes a subset of files, Critics ensure no vulnerabilities were missed across the entire project, and the Summarizer provides a consolidated report of all findings.

Quick Start

Use the doing-a-simple-two-stage-fanout skill to analyze the corpus located at /path/to/large/dataset.txt with a 'some effort' level of analysis.

Frequently Asked Questions about doing-a-simple-two-stage-fanout

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

FAQPage Schema
How do I analyze large datasets that exceed a single agent's context window?

To analyze massive text, code, or data corpuses that exceed single agent context limits, use a two-stage fan-out pipeline that divides the corpus into parallel Worker segments, followed by Critic review and Summarizer synthesis.

What is the best way to ensure quality assurance during parallel processing of a large codebase?

Quality assurance during parallel processing is achieved by employing Critic agents to review Worker outputs, identifying gaps, inconsistencies, and cross-segment patterns before final synthesis.

Can I use distributed computing to find security vulnerabilities across a multi-gigabyte codebase?

Yes, distributed computing orchestrates multiple Worker agents to analyze subsets of files in parallel. Critics then ensure no vulnerabilities were missed across the project, consolidating all findings into a final report.

How does agent orchestration handle failure recovery in large scale analysis?

Agent orchestration handles failure recovery in large scale analysis by incorporating task tracking and recovery mechanisms throughout the Worker, Critic, and Summarizer pipeline stages.

What are the limitations of using a two-stage fanout approach for data processing?

The two-stage fanout approach for data processing requires segmenting the corpus, meaning it is limited to scenarios where dividing the dataset does not break critical cross-segment context dependencies entirely.