ag-learn

Summarize and persist agent session learnings into a durable log.

21|Updated Jan 3, 2026

Search Tags:#audit #llm #improvement #reflection #sessions #workflows #learn

Ready to run

npx skills add https://github.com/kevinslin/skills --skill ag-learn

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ag-learn
Source: https://github.com/kevinslin/skills/tree/main/active/ag-learn
Command: npx skills add https://github.com/kevinslin/skills --skill ag-learn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Consolidates insights from the current and prior agent sessions into a durable learnings log to drive continuous improvement and reduce repeated mistakes.

Core Features & Use Cases

Time-bounded reviews across single or multiple sessions to surface mistakes, uncertainties, and optimization opportunities.
Routing learnings to skills, AGENTS.md, repository docs, or personal workflows to guide future work.
Persisting syntheses into a durable learnings log for audit and replay.

Quick Start

Ask me to run the default learn cycle on the current session to generate a concise improvement report.

Frequently Asked Questions about ag-learn

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

FAQPage Schema

How do I consolidate learnings from LLM agent sessions into a durable log?▼

Consolidating learnings from LLM agent sessions involves summarizing insights and persisting them into a durable log to drive continuous improvement and reduce repeated mistakes across workflows.

How does session reflection identify mistakes and optimization opportunities?▼

Session reflection identifies mistakes and optimization opportunities by applying time-bounded reviews across single or multiple sessions to surface uncertainties and route learnings to skills or repository docs.

What is the best way to audit agent workflows for continuous improvement?▼

Auditing agent workflows for continuous improvement requires summarizing session data, persisting syntheses into a durable learnings log, and generating rollup reports to guide future work.

Can I apply learnings from previous sessions to improve future LLM outputs?▼

Yes, you can apply learnings from previous sessions to single-session or multi-session windows to identify past mistakes and route insights to personal workflows for optimizing future LLM outputs.

Does session learning require specific file naming conventions for audit logs?▼

Yes, ensuring output conforms to required file naming and rollup reporting is necessary when persisting syntheses into a durable learnings log for audit and replay purposes.

When should I not use automated session reflection for workflow improvement?▼

Automated session reflection is not ideal when lacking sufficient session data, as time-bounded reviews require multiple sessions to effectively surface mistakes, uncertainties, and optimization opportunities.