continuous-learning

Capture reusable patterns from Claude Code sessions and save them as learned skills.

Updated Mar 26, 2026
One-click install
npx skills add https://github.com/cescrafli/compyrasion --skill continuous-learning-cescrafli
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: continuous-learning
Source: https://github.com/cescrafli/compyrasion/tree/main/skills/continuous-learning
Command: npx skills add https://github.com/cescrafli/compyrasion --skill continuous-learning-cescrafli

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq, and includes scripts (resource) components.

What problem does it solve?

Automatically captures reusable patterns from Claude Code sessions and persists them as learned skills for future reuse, reducing manual pattern discovery and speeding setup of productive workflows.

Core Features & Use Cases

  • Identifies actionable patterns from Claude Code transcripts (e.g., errors, corrections, workarounds) and saves them into a centralized learned-skills directory.
  • Provides configurable thresholds and directories via config.json to tailor extraction and storage behavior.
  • Supports Stop-hook integration to run at session end, ensuring non-disruptive pattern mining across many sessions.

Quick Start

Run the continuous-learning stop-hook evaluator to scan the current Claude Code session and save detected patterns to ~/.claude/skills/learned/.

Frequently Asked Questions about continuous-learning

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

FAQPage Schema
How do I automatically extract reusable patterns from Claude Code sessions?

You can automatically extract reusable patterns from Claude Code sessions by using a stop-hook evaluator that identifies error resolutions, user corrections, and workarounds, saving them directly to a learned skills directory for future reuse.

What types of session patterns can be captured and saved for continuous learning?

The continuous learning process captures specific session patterns including error_resolution, user_corrections, workarounds, debugging_techniques, and project_specific knowledge, persisting them as learned skills.

How do I configure minimum session length and extraction thresholds for pattern mining?

You can configure pattern mining behavior by setting min_session_length, extraction_threshold, and learned_skills_path variables in the config.json file to control when and how session evaluations trigger.

Do I need jq installed to run Claude Code session evaluations?

Yes, jq is a required dependency for parsing and extracting session data during the stop-hook evaluation process to ensure accurate pattern identification and storage.

Can I use a stop-hook to run non-disruptive pattern mining at the end of a session?

Yes, integrating a stop-hook allows pattern mining to run automatically at session end, ensuring non-disruptive extraction of debugging techniques and workarounds across many Claude Code sessions.

What is the default directory path for storing learned skills from Claude Code?

The default directory for storing learned skills is the local path ~/.claude/skills/learned/, where all extracted session patterns are persisted for future workflow reuse.