continuous-learning

To automatically generate reusable skills from code session transcripts, configure a continuous-learning stop hook. It analyzes completed sessions to extract problem-solving patterns and error resolutions, saving them as new skills for future use.

The best way to capture user-correction strategies is through automatic session analysis. This process detects repeatable solutions and codifies them into actionable skills, making your novel debugging techniques available for future similar problems.

To set up automatic pattern detection for Claude Code sessions, configure the continuous-learning skill to run as a stop hook in your settings.json. This triggers session evaluation automatically upon meeting a minimum length threshold.

Extracting reusable patterns requires sessions meeting a minimum length threshold. Short debugging sessions may not trigger the automatic evaluation needed to identify and codify problem-solving techniques into newly generated skills.

Session analysis for skill generation requires no external dependencies. It relies solely on internal pattern detection logic applied to your session transcripts to identify and save actionable insights as learned skills.

You need to codify problem-solving techniques into automated skills when you want to reuse solutions for similar problems. This applies to scenarios requiring pattern recognition for error resolutions and user-correction strategies extracted from session transcripts.