What problem does it solve? Valuable debugging techniques, workarounds, and repeated command patterns discovered during Claude Code sessions are lost when the session ends, forcing you to rediscover them in future sessions. ## Core Features & Use Cases - Session-End Pattern Extraction: A Stop hook analyzes the transcript with claude -p and saves reusable patterns as discoverable Agent Skills under the project's .claude/skills/ directory. - CLI Suggestions: Detects repeated shell command patterns and proposes new subcommands for the project's own CLI, written to .claude/cli-suggestions/_pending/ for review. - Curator Lifecycle Management: Tracks usage telemetry, moves idle skills through active/stale/archived states, deduplicates near-identical skills, and snapshots backups before every run. - Use Case: After a long debugging session where you fixed a tricky Django ORM issue, the hook automatically saves that resolution as a learned skill so the next session in the same repo can apply it immediately. ## Quick Start Set up the continuous-learning Stop hook so reusable patterns from my Claude Code sessions are saved as skills in this project.
