index-knowledge

Generate hierarchical AGENTS.md knowledge bases for code repositories.

Updated Mar 25, 2026
One-click install
npx skills add https://github.com/involvex/skills --skill index-knowledge-involvex
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: index-knowledge
Source: https://github.com/involvex/skills/tree/main/skills/index-knowledge
Command: npx skills add https://github.com/involvex/skills --skill index-knowledge-involvex

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates creating a concise, hierarchical AGENTS.md knowledge base for code repositories so maintainers can quickly discover structure, conventions, and hotspots without manual documentation effort.

Core Features & Use Cases

  • Automated Discovery: Parallel explore agents, bash structural analysis, and optional LSP codemaps to find entry points, large files, and module boundaries.
  • Scoring & Placement: A configurable scoring matrix determines where to create root and subdirectory AGENTS.md files to avoid redundancy and surface complexity.
  • Generation & Review: Produces root and per-directory AGENTS.md in parallel, applies deduplication and size gates, and emits a final report listing created/updated files.
  • Use Case: Ideal for onboarding, codebase audits, and maintaining documentation in monorepos where directory complexity and conventions vary.

Quick Start

Generate an AGENTS.md hierarchy for the current repository, scoring directories, creating a root AGENTS.md and selective subdirectory AGENTS.md, and produce a final summary report.

Frequently Asked Questions about index-knowledge

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

FAQPage Schema
How do I generate an AGENTS.md knowledge base for a complex code repository?

Generate an AGENTS.md knowledge base by analyzing project structure, deploying parallel explore agents, and using LSP codemaps to detect directories, score complexity, and decide root and subdirectory file placement.

What is the best way to create hierarchical documentation for a monorepo?

Create hierarchical documentation for a monorepo by applying a configurable scoring matrix to directory complexity, detecting module boundaries, and generating concise, non-redundant AGENTS.md files across the project structure.

How does scoring complexity help decide where to place AGENTS.md files?

Scoring complexity evaluates directory structure and file sizes to determine AGENTS.md placement, ensuring documentation surfaces high-complexity hotspots while avoiding redundant files in simple subdirectories.

Can I use LSP codemaps to find entry points and module boundaries in my codebase?

LSP codemaps are used alongside parallel explore agents and bash structural analysis to automatically discover entry points, large files, and module boundaries within your code repository.

Does generating AGENTS.md files work for both monorepos and single-package projects?

Generating AGENTS.md files applies across both monorepos and single-package projects, analyzing structural complexity to produce root and selective subdirectory documentation with a final summary report.

What happens when AGENTS.md generation encounters redundant directory structures?

When encountering redundant directory structures, generation applies deduplication and size gates to ensure produced AGENTS.md files remain concise and non-redundant across the repository hierarchy.