self-learn

Analyze recent session learnings and update MEMORY.md with verified insights.

Updated Mar 16, 2026
One-click install
npx skills add https://github.com/shihwesley/shihwesleys-harness --skill self-learn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: self-learn
Source: https://github.com/shihwesley/shihwesleys-harness/tree/main/skills/self-learn
Command: npx skills add https://github.com/shihwesley/shihwesleys-harness --skill self-learn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Trigger self-learning synthesis to analyze recent session learnings and update MEMORY.md with verified insights.

Core Features & Use Cases

  • Analyze recent learnings from sessions and cluster recurring patterns.
  • Update MEMORY.md with actionable insights, pitfalls, and recommendations.
  • Run on demand via /self-learn or with an optional --force flag.

Quick Start

Initiate the self-learning synthesis to analyze recent learnings and update MEMORY.md.

Frequently Asked Questions about self-learn

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

FAQPage Schema
How do I consolidate session learnings into actionable insights?

Consolidate session learnings by triggering a self-learning synthesis that analyzes recent sessions, clusters recurring patterns, and updates your memory file with verified insights. This automation runs on demand via a simple command.

How does self-learning synthesis update MEMORY.md with new patterns?

Self-learning synthesis updates MEMORY.md by clustering recurring patterns from recent learnings and writing actionable insights, pitfalls, and recommendations directly into the file. It uses a threshold gate to ensure sufficient data before running.

When do I need to automate learning synthesis across sessions?

Automate learning synthesis across sessions when you want to review recurring patterns, consolidate verified insights, or prepare memory before starting new projects. A threshold requires at least 5 new learnings since the last synthesis.

Can I force learning synthesis to run with fewer than 5 new learnings?

Yes, you can force learning synthesis to run with fewer than 5 new learnings by using the optional --force flag. This bypasses the default threshold gate and triggers the synthesis process immediately.

Does Claude Haiku work for automating self-learning pattern clustering?

Yes, Claude Haiku works for automating self-learning pattern clustering. The synthesis process integrates with your memory file and runs via Claude Haiku to analyze recent learnings and extract actionable insights.

What are the limitations of threshold-gated learning synthesis?

The limitation of threshold-gated learning synthesis is that it requires at least 5 new learnings since the last synthesis to run automatically. You must use the --force flag to bypass this constraint when you have insufficient new data.