rekal-init

Scan a project codebase and store architecture, dependencies, and workflows as durable memories.

52|4|Updated Apr 11, 2026
One-click install
npx skills add https://github.com/janbjorge/rekal --skill rekal-init
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: rekal-init
Source: https://github.com/janbjorge/rekal/tree/main/skills/rekal-init
Command: npx skills add https://github.com/janbjorge/rekal --skill rekal-init

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

New or existing Claude Code agents repeatedly ask for project details, causing redundant conversations. This skill scans the codebase, extracts architecture, dependencies, conventions, and workflows, and stores them as durable memories so the agent starts with full context.

Core Features & Use Cases

  • Automated Project Scanning: Reads key documentation, config files, and entry points to capture essential knowledge.
  • Selective Extraction: Filters information to retain only high‑value facts, preferences, and procedures.
  • Deduplication & Supersession: Searches existing memories to avoid duplicates and updates outdated entries.
  • Safety Checks: Confirms before overwriting existing project memories and respects privacy constraints.
  • Large‑Codebase Handling: Allows focused scanning of monorepo sections and incremental progress reporting.

Quick Start

Ask the assistant to run “/rekal-init” to scan the repository and populate rekal memory with project context.

Frequently Asked Questions about rekal-init

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

FAQPage Schema
How do I bootstrap project context for Claude Code agents to avoid redundant conversations?

To bootstrap project context, scan the codebase to extract architecture, dependencies, and workflows, then store them as durable memories. This gives agents full session context without repeatedly asking for project details.

What is the best way to scan a large monorepo and extract codebase architecture for AI memory?

The best way to scan a large monorepo is using focused, incremental scanning on specific sections. This selectively extracts high-value facts and workflows while reporting progress, ensuring comprehensive memory without overwhelming the system.

Can I update existing project memories without creating duplicates when scanning my codebase?

Yes, you can update existing memories without duplicates. The scanning process searches existing memory stores to avoid duplicates and updates outdated entries, with safety checks that confirm before overwriting existing project memories.

Do I need read access to project files to extract dependencies and workflows for agent context?

Yes, you need read access to project files. Extracting dependencies and workflows requires utilizing allowed tools for globbing, searching, and memory store operations to capture essential knowledge from documentation and config files.

What limitations exist when bootstrapping memory for a new or existing project codebase?

Limitations include respecting privacy constraints during selective extraction and requiring safety checks before overwriting existing memories. The process filters to retain only high-value facts, which may require multiple passes for very large codebases.