dream

Capture user preferences and project rules into persistent memory files.

38|10|Updated Oct 29, 2025
One-click install
npx skills add https://github.com/maxvaega/awesome-skills --skill dream-maxvaega
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dream
Source: https://github.com/maxvaega/awesome-skills/tree/main/dream
Command: npx skills add https://github.com/maxvaega/awesome-skills --skill dream-maxvaega

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq, and includes scripts (resource) components.

What problem does it solve?

This skill solves the problem of AI agents forgetting user preferences, project-specific rules, and feedback over time, ensuring that your AI assistant becomes more personalized and effective with every session.

Core Features & Use Cases

  • Durable Memory: Automatically captures user preferences, corrections, and standing rules into project-level memory files.
  • Self-Improvement: Periodically reviews session history to propose surgical improvements to your skills and CLAUDE.md files.
  • REM Consolidation: Performs weekly maintenance to merge overlapping memories, archive stale information, and keep your context window lean.

Quick Start

Run the dream command in your terminal to capture the current session learnings and store them as durable project memories.

Frequently Asked Questions about dream

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

FAQPage Schema
How do I make Claude Code remember my project rules and preferences across sessions?

You can persist project rules by using a memory loop skill that captures feedback and corrections into durable memory files. It automatically consolidates session learnings so Claude Code retains context across sessions.

What is automated agent self-improvement in Claude Code?

Automated agent self-improvement is the process of periodically analyzing session history to propose surgical updates to skills and configuration files. This mechanism allows Claude Code to iteratively refine its behavior based on past interactions.

How do I prevent context window overload when storing Claude Code memories?

You can prevent context window overload by running weekly memory consolidation maintenance. This process merges overlapping memories, archives stale information, and prunes redundant data to keep your context lean and efficient.

Does Claude Code memory automation require any specific dependencies?

Yes, Claude Code memory automation requires the jq command-line tool for efficient transcript processing. It also relies on standard Unix utilities for cron-based automation and file management.

How do I automatically capture session learnings in Claude Code?

You can automatically capture session learnings by running a terminal command that triggers the memory loop. This analyzes your current session transcripts and extracts durable user preferences and project rules into a persistent store.