learning-loop

Log newly discovered learnings to .claude/learning.yaml with timestamps.

306|17|Updated Sep 5, 2025
One-click install
npx skills add https://github.com/PackmindHub/packmind --skill learning-loop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: learning-loop
Source: https://github.com/PackmindHub/packmind/tree/main/.claude/skills/learning-loop
Command: npx skills add https://github.com/PackmindHub/packmind --skill learning-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Capture newly discovered learnings during code exploration, user interaction, or tool discovery and log them for future Claude sessions.

Core Features & Use Cases

  • Log new learnings to .claude/learning.yaml with a timestamp to preserve context.
  • Append concise summaries to guide future sessions and avoid repeating work.
  • Use Case: While exploring a complex codebase, capture undocumented behaviors or domain knowledge for later review.

Quick Start

Use the learning-loop skill's script to log a new learning: .claude/skills/learning-loop/scripts/log-learning.sh "Observed undocumented behavior: API returns 201 on create"

Frequently Asked Questions about learning-loop

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

FAQPage Schema
How do I capture and store learnings from code exploration in Claude sessions?

Logging learnings preserves undocumented behaviors, hidden configurations, and domain knowledge discovered during code exploration. Use the learning-loop Skill to append entries to .claude/learning.yaml with timestamps, creating a persistent knowledge base that informs future Claude sessions and prevents repeating discovery work.

What's the best way to document undocumented API behavior for reuse?

Write concise summaries of undocumented behavior to .claude/learning.yaml using the log-learning.sh script. Each entry captures the discovery context and timestamp, making it retrievable and actionable in subsequent sessions without manual re-exploration.

Can I use learning-loop to build a knowledge base from tool discovery?

Yes. Learning-loop logs new discoveries about tools, configurations, and domain-specific knowledge directly to a YAML file. This creates a cumulative knowledge base that Claude can reference in future sessions, eliminating duplicate discovery and building institutional memory.

How does logging learnings improve efficiency across multiple Claude sessions?

By writing timestamped entries to .claude/learning.yaml, learning-loop ensures insights discovered in one session remain accessible in the next. This avoids re-exploration, accelerates problem-solving, and maintains continuity of domain knowledge across your Claude interactions.

What format does learning-loop use to store captured knowledge?

Learning-loop stores all entries in YAML format within .claude/learning.yaml. This human-readable, structured format preserves timestamps and summaries, making learnings easy to review, query, and reference across sessions without parsing overhead.