learn

Capture user preferences and session patterns into domain-specific learning files.

8|1|Updated Apr 3, 2026
One-click install
npx skills add https://github.com/agentflock/myna --skill learn-agentflock
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: learn
Source: https://github.com/agentflock/myna/tree/main/skills/learn
Command: npx skills add https://github.com/agentflock/myna --skill learn-agentflock

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Managing emergent memory in long-running AI assistants requires capturing user preferences and observed patterns, while avoiding noise and entity-specific data.

Core Features & Use Cases

  • Capture learnings from conversations and route them to domain-specific files under _meta/learnings/.
  • Manage Active vs Proposed entries, reflect on patterns, forget or promote rules, and perform negotiation when needed.
  • Support periodic reflection to surface user patterns for refinement.

Quick Start

Tell Myna to remember that you prefer concise meeting summaries.

Frequently Asked Questions about learn

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

FAQPage Schema
How do I save user preferences so my AI assistant remembers them across sessions?

To save user preferences, this Skill captures conversation patterns and routes them as learnings into domain-specific files under _meta/learnings/, ensuring your AI retains behavioral rules across sessions. It organizes captured preferences into domains like general, email, meetings, and people.

What is the best way to manage emergent memory without storing noisy entity-specific data in an AI?

Managing emergent memory effectively involves capturing user preferences and observed patterns while filtering out noise. This Skill organizes patterns into Active and Proposed entries, using a negotiation flow to promote only validated rules, preventing unverified data from cluttering long-term memory.

How do I promote a proposed learning to an active rule for my AI?

To promote a proposed learning to an active rule, the Skill uses a built-in negotiation flow that evaluates Proposed entries before moving them to Active status. This ensures only verified patterns shape your AI's behavior in specific domains like email or meetings.

Can I reflect on and forget previously learned AI patterns?

Yes, you can reflect on and forget learned patterns. The Skill supports periodic reflection to surface user patterns for refinement, and includes a forget function to remove outdated or incorrect learnings from the domain-specific memory files.

Do I need a vault to store captured AI learnings and preferences?

Yes, a vault is required to store captured learnings and preferences. The Skill requires vault-backed writes to save and organize session patterns into the _meta/learnings/ directory structure across various domains.