project-indexing

Compress codebases into a 3K SHANNON_INDEX with structured metadata.

2|Updated Sep 30, 2025

Search Tags:#indexing #codebase #compression #serena #token reduction

Ready to run

npx skills add https://github.com/krzemienski/shannon-framework --skill project-indexing

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: project-indexing
Source: https://github.com/krzemienski/shannon-framework/tree/main/skills/project-indexing
Command: npx skills add https://github.com/krzemienski/shannon-framework --skill project-indexing

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Glob, Read, Grep, Serena, and includes references (resource) components.

What problem does it solve?

Shannon codebase compression producing 94% token reduction and a structured index for fast context.

Core Features & Use Cases

94% token compression (e.g., 58K → ~3K)
Core modules, tech stack, dependencies, recent changes
Serena persistence for cross-session access
Rapid queries with structured index

Quick Start

Generate a Shannon index for a large codebase and begin querying context from the index.

Frequently Asked Questions about project-indexing

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

FAQPage Schema

How much can codebase compression reduce token usage?▼

Codebase compression using Shannon indexing achieves 94% token reduction. For example, a 58K codebase compresses to approximately 3K, enabling efficient context window usage during code exploration and multi-agent coordination.

How do I compress a large codebase into a searchable index?▼

Generate a Shannon index by running project-indexing on your codebase. The Skill produces a structured 3K index containing core modules, tech stack, and dependencies, which persists in Serena for fast cross-session retrieval and agent onboarding.

Can I regenerate the index when my codebase changes?▼

Yes. The Skill supports index regeneration on codebase changes, allowing you to keep the compressed index synchronized with your project's current state while maintaining Serena persistence.

What's the best way to onboard multiple agents into a codebase quickly?▼

Use the Shannon index for rapid agent onboarding. The structured metadata index enables fast context loading and multi-agent coordination by reducing token waste from unindexed code exploration.

Does this work for switching between different projects?▼

Yes. Serena persistence stores the compressed index across sessions, making context switching between codebases efficient by eliminating redundant compression work and providing instant structured metadata retrieval.

What dependencies does codebase indexing require?▼

The Skill requires Glob for file traversal, Read for content extraction, Grep for pattern matching, and Serena for persistent index storage and cross-session retrieval of compressed codebase metadata.