How does Session Learning identify relevant past insights?

It uses a keyword matching algorithm that compares your current task prompt against the tags and technical terms stored in your categorized learning YAML files.

Will this impact the performance of my Claude sessions?

No, the skill follows a fail-safe principle. If extraction or injection fails or takes too long, the system continues the session normally without blocking your workflow.

Does this replace DISCOVERIES.md or PATTERNS.md files?

No, it is designed to complement them. While those files are typically for human-readable milestones, this skill handles granular, reusable technical insights for the AI agent to use automatically.

Where are the learnings stored?

Insights are stored locally in your project directory under .claude/data/learnings/ in structured YAML files like errors.yaml and architecture.yaml.

Can I manually manage or view the collected learnings?

Yes, you can use the /amplihack:learnings command to search, display, manually add, or view statistics for all collected insights across your project.

Session Learning

Name: Session Learning
Author: rysweet

byrysweet

•

Productivity & Workflow

Captures and reapplies technical insights across coding sessions to build a persistent memory of fixes, workflows, and architectural decisions.

The Session Learning skill transforms Claude into a self-improving development agent by systematically extracting structured insights from session transcripts. By storing learnings in categorized YAML files—including errors, workflows, tools, and architecture—it ensures that hard-won solutions and tribal knowledge are preserved. At the start of new tasks, the skill automatically injects relevant past context based on keyword matching, significantly reducing redundant troubleshooting and accelerating the development lifecycle within the Amplihack framework.

Key Features

01Keyword-based context injection at the start of new sessions for relevant guidance.

02Automatic insight extraction from session transcripts upon session exit.

03Structured YAML storage for errors, workflows, tools, and architecture.

0416 GitHub stars

05Fail-safe design that ensures learning processes never block session startup or shutdown.

06Built-in /amplihack:learnings command for manual search and insight management.

Use Cases

01Automating the documentation of 'gotchas' discovered during intense debugging sessions without manual note-taking.

02Onboarding to a project by reviewing historical design decisions and debugging strategies stored in YAML.

03Preventing the recurrence of complex circular dependency or environment-specific configuration errors.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rysweet/amplihack session-learning

For use in Claude.ai and ChatGPT

Download Skill