pi-learn

Store and organize peer representations, observations, conclusions, and summaries across global and project scopes.

4|2|Updated Mar 16, 2026
One-click install
npx skills add https://github.com/0xKobold/0xkobolds --skill pi-learn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pi-learn
Source: https://github.com/0xKobold/0xkobolds/tree/main/packages/pi-learn
Command: npx skills add https://github.com/0xKobold/0xkobolds --skill pi-learn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

pi-learn provides a robust memory infrastructure for pi agents, enabling persistent memory, peer representations, and contextual reasoning across multiple sessions and projects.

Core Features & Use Cases

  • Hybrid memory architecture with global (user) and project (local) scopes to keep personal traits and project patterns separate.
  • Memory, context assembly, and reasoning capabilities including a dreaming mechanism for creative insights.
  • Retention and pruning policies to manage data lifecycle and privacy across workspaces.

Quick Start

Install pi-learn into your pi workflow and start the extension to enable persistent agent memory across projects.

Frequently Asked Questions about pi-learn

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

FAQPage Schema
How do I add persistent memory to AI agents across multiple projects?

You add persistent memory by implementing a hybrid memory store that saves observations, conclusions, and peer representations across global and project scopes, maintaining context across multiple sessions.

What is hybrid memory architecture for AI agents?

Hybrid memory architecture separates agent memory into global scopes for user traits and project scopes for local patterns. It supports context assembly, reasoning, and retention policies to manage data lifecycle and privacy across workspaces.

How do agents maintain context across different sessions and projects?

Agents maintain context by storing peer representations and summaries in a memory infrastructure that applies previous observations to current reasoning tasks across different sessions and projects.

Can I manage memory retention and data privacy for my AI agents?

Yes, you can manage data privacy by applying retention and pruning policies to control the data lifecycle, organizing what observations and conclusions persist across different workspaces.

What is the dreaming mechanism in agent reasoning?

The dreaming mechanism is a memory infrastructure feature that processes stored observations and conclusions to generate creative insights, operating alongside context assembly to support advanced reasoning.

Does pi-learn work without external dependencies?

Yes, pi-learn operates without external dependencies, providing tools, context assembly, and retention policies directly within your agent workflow to enable persistent memory.